Skip to main content

Full text of "Science progress"

See other formats














VOL. V. 














All rights reserved 





African Grass Fires and their Effects. By G. F. Scott Elliott, M.A. 77 

The Past, Present, and Future Water Supply of London. By E. 

Frankland, F.R.S. 163 

Gold Extraction Processes. By T. K. Rose, D.Sc, Assistant Assayer 

of the Royal Mint ------- 484 


The Morphology of the Mollusca. By. W. Garstang, M.A., Fellow 

of Lincoln College, Oxford 38 

The Present Position of the Cell Theory, By G. C. Bourne, M.A., 

Fellow of New College, Oxford - - -94,227,304 

Some Recent Memoirs upon Oligochfeta. By F. E. Beddard, 

M.A., F.R.S. 19° 


Pre-historic Man in the Eastern Mediterranean. By J. L. Myers, 

Fellow of Magdalen College, Oxford - - 335 

Selection in Man. By John Beddoe, M.D., LL.D., F.R.S. - - 384 


On Recent Advances in Vegetable Cytology. By J. Bretland Farmer, 
M.A., Professor of Botany in the Royal College of Science, 
London - ---------22 

The Reserve Materials of Plants [Concluded). By J. Reynolds 
Green, M.A., F.R.S., Professor of Botany to the Pharma- 
ceutical Society, London ------ 60 

The Stelar Theory : A History and a Criticism. By A. G. Tansley, 

B.A. i33j 215 

Ferns : Aposporous and Apogamous. By C. T. Druery, F.L.S. - 242 

Insular Floras. By W. Botting Hemsley, F.R.S.- - - 286,374 


The General Bearings of Magnetic Observations. By Captain Ettrick 
W. Creak, R.N., F.R.S., Superintendent of Compasses to 
the Admiralty - - - - - - - - - 81 

Solid Solutions. By James Walker, Ph.D., D.Sc, Professor of 

Chemistry in University College, Dundee - - - - 121 

2 /Sf& 



Notes on Atomic Weights. By Alexander Scott, M.A., Jacksonian 

Demonstrator in the University of Cambridge - 202 

The Growth of our Knowledge of Helium. By J. Norman Lockyer, 

C.B., F.R.S. --------- 249 

Light and Electrification. By Oliver Lodge, F.R.S., Professor of 

Physics in University College, Liverpool - - - - 417 

Recent Values of the Magnetic Elements at the Principal Magnetic 
Observatories of the World. By Charles Chree, M.A., 
Superintendent of Kew Observatory - - - - - 499 



The Graptolites. By J. E. Marr, M.A., F.R.S., Fellow of St. John's 

College, Cambridge -------- 360 

Recent Discoveries in Avian Palaeontology. By C. W. Andrews - 398 
An Extinct Plant of Doubtful Affinity. By A. C. Seward, M.A., 

F.G.S., University Lecturer in Botany, Cambridge - 428 

The Work of the Portuguese Geological Survey. By Philip Lake, 

M. A., St. John's College, Cambridge- - 439 

Petrology in America. By Alfred Harker, M. A., Fellow of St. John's 

College, Cambridge- ------- 459 


The Hereditary Transmission of Micro-organisms. By G. A. Buck- 
master, M.D., Lecturer on Physiology at St. George's Hospital, 

London ---------- 324 


Ludwig and Modern Physiology. By J. Burdon Sanderson, M.D., 
F.R.S., Regius Professor of Physiology in the University of 
Oxford ---------- 1 

On Some Applications of the Theory of Osmotic Pressures to 
Physiological Problems (Part II.). By E. H. Starling, M.D., 
Lecturer on Physiology at Guy's Hospital, London - - 151 

Iodine in the Animal Organism. By W. D. Halliburton, M.D., 

F.R.S., Professor of Physiology in King's College, London - 454 

Notices of Books, - - - - - - 1, xi, xxi, xxxi, xli 

Titles of Chemical Papers, - iv, xv, xxv, xxxm, xlii, xlvii 



Andrews, C W. Recent Discoveries in Avian Palaeontology - 398 

Beddard, F. E. Some Recent Memoirs upon Oligochaeta - 190 

Beddoe, John. Selection in Man - 384 

Bourne, G. C. The Present Position of the Cell Theory 94, 227, 304 
Buckmaster, G. A. The Hereditary Transmission of Micro- 
organisms - - - - - - - ~3 2 4 

Chree, Charles. Recent Values of the Magnetic Elements - - 499 
Creak, Captain Ettrick. The General Bearings of Magnetic Observa- 
tions ___-__--- 81 
Druery, C. T. Ferns : Aposporous and Apogamous - - 242 
Farmer, J. B. On Recent Advances in Vegetable Cytology - - 22 
Frankland, E. The Past, Present and Future Water Supply of 

London ------- 163 

Garstang, W. The Morphology of the Mollusca - 38 

Green, J. Reynolds. The Reserve Materials of Plants - 60 

Halliburton, W. D. Iodine in the Animal Organism - 454 

Harker, Alfred. Petrology in America - - 459 
Hemsley, W. Botting. Insular Floras - - - 286, 374 

Lake, Philip. The Work of the Portuguese Geological Survey 439 

Lockyer, J. Norman. The Growth of our Knowledge of Helium - 249 

Lodge, Oliver. Light and Electrification - - - - - 417 

Marr, J. E. The Graptolites ------ 360 

Myers, J. L. Prehistoric Man in the Eastern Mediterranean - 335 

Rose, T. R. Gold Extraction Processes ----- 484 

Sanderson, J. Burdon. Ludwig and Modern Physiology - - 1 

Scott, Alexander. Notes on Atomic Weights ... - 202 

Scott-Elliott, G. F. African Grass Fires and their Effects - - 77 

Seward, A. C. An Extinct Plant of Doubtful Affinity - 4 2 8 
Starling, E. H. On some Applications of the Theory of Osmotic 

Pressures to Physiological Problems - - - - - 151 

Tansley, A. G. The Stellar Theory - 133,215 

Walker, James. Solid Solutions - - - - - - - 121 


Btimce progress 

ku UlRARYbj 

- ^7- 

a. ' CV 


No. 25. 

March, 1896. 

Vol. V. 



THE death of any discoverer — of any one who has 
added largely to the sum of human knowledge — 
affords a reason for inquiring what his work was and how 
he accomplished it. This inquiry has interest even when 
the work has been completed in a few years and has been 
limited to a single line of investigation — much more when 
the life has been associated with the origin and develop- 
ment of a new science and has extended over half a 

The Science of Physiology as we know it came into 
existence fifty years ago with the beginning of the active life 
of Ludwig, in the same sense that the other great branch of 
Biology, the Science of Living Beings (Ontology), as we 
now know it, came into existence with the appearance of the 
" Origin of Species ". In the order of time Physiology had 
the advantage, for the new Physiology was accepted some 
ten years before the Darwinian epoch. Notwithstanding, the 
content of the science is relatively so unfamiliar, that before 
entering: on the discussion of the life and work of the man 
who, as I shall endeavour to show, had a larger share in 
founding it than any of his contemporaries, it is necessary 
to define its limits and its relations to other branches of 

1 Founded upon a lecture delivered at the Royal Institution, Jan. 
24, 1896. 


The word Physiology has in modern times changed its 
meaning. It once comprehended the whole knowledge of 
Nature. Now it is the name for one of the two Divisions 
of the Science of Life. In the progress of investigation 
the study of that Science has inevitably divided itself into 
two : Ontology, the Science of Living Beings ; Physiology, 
the Science of Living Processes, and thus, inasmuch as 
Life consist in processes, of Life itself. Both strive to 
understand the complicated relations and endless varieties 
which present themselves in living Nature, but by different 
methods. Both refer to general principles, but they are of 
a different nature. 

To the Ontologist, the student of Living Beings, Plants 
or Animals, the great fact of Evolution, namely, that from 
the simplest beginning our own organism, no less than that 
of every animal and plant with its infinite complication of 
parts and powers, unfolds the plan of its existence — taken 
with the observation that that small beginning was, in all 
excepting the lowest forms, itself derived from two parents, 
equally from each — is the basis from which his study and 
knowledge of the world of living beings takes its departure. 
For on these two facts — Evolution and Descent — the ex- 
plorer of the forms, distribution and habits of animals and 
plants has, since the Darwinian epoch, relied with an ever- 
increasing certainty, and has found in them the explanation 
of every phenomenon, the solution of every problem relating 
to the subject of his inquiry. Nor could he wish for a more 
secure basis. Whatever doubts or misgivings exist in the 
minds of " non-biologists " in relation to it, may be attributed 
partly to the association with the doctrine of Evolution of 
questions which the true naturalist regards as transcen- 
dental ; partly to the perversion or weakening of meaning 
which the term has suffered in consequence of its introduc- 
tion into the language of common life, and particularly to 
the habit of applying it to any kind of progress or improve- 
ment, anything which from small beginnings gradually 
increases. But, provided that we limit the term to its 
original sense — the Evolution of a living being from its 
germ by a continuous, not a gradual process — there is no 


conception which is more free from doubt either as to its 
meaning or reality. It is inseparable from that of Life 
itself, which is but the unfolding of a predestined harmony, 
of a prearranged consensus and synergy of parts. 

The other branch of Biology, that with which Ludwig's 
name is associated, deals with the same facts in a different 
way. While Ontology regards animals and plants as in- 
dividuals and in relation to other individuals, Physiology 
considers the processes themselves of which life is a complex. 
This is the most obvious distinction, but it is subordinate to 
the fundamental one, namely, that while Ontology has for 
its basis laws which are in force only in its own province, 
those of Evolution, Descent, and Adaptation, we Physiolo- 
gists, while accepting these as true, found nothing upon them, 
using them only for euristic purposes, i.e., as guides to dis- 
covery, not for the purpose of explanation. Purposive Adapta- 
tion, for example, serves as a clue, by which we are constantly 
guided in our exploration of the tangled labyrinth of vital 
processes. But when it becomes our business to explain 
these processes — to say how they are brought about — we 
refer them not to biological principles of any kind, but to 
the Universal Laws of Nature. Hence it happens that 
with reference to each of these processes, our inquiry is 
rather how it occurs than why it occurs. 

It has been well said that the Natural Sciences are the 
children of necessity. Just as the other Natural Sciences 
owed their origin to the necessity of acquiring that control 
over the forces of Nature without which life would 
scarcely be worth living, so Physiology arose out of human 
suffering and the necessity of relieving it. It sprang indeed 
out of Pathology. It was suffering that led us to know, as 
regards our own bodies, that we had internal as well as 
external organs, and probably one of the first generalisa- 
tions which arose out of this knowledge was, that " if one 
member suffer all the members suffer with it " — that all 
work together for the good of the whole. In earlier times 
the good which was thus indicated was associated in men's 
minds with human welfare exclusively. But it was 
eventually seen that Nature has no less consideration for 


the welfare of those of her products which to us seem 
hideous or mischievous, than for those which we regard 
as most useful to man or most deserving of his admiration. 
It thus became apparent that the good in question could not 
be human exclusively, but as regards each animal its oivn 
good — and that in the organised world the existence and 
life of every species is brought into subordination to one 
purpose — its own success in the struggle for existence. 1 

From what has preceded it may be readily understood 
that in Physiology, Adaptation takes a more prominent 
place than Evolution or Descent. In the prescientific 
period adaptation was everything. The observation that 
any structure or arrangement exhibited marks of adaptation 
to a useful purpose was accepted not merely as a guide in 
research, but as a full and final explanation. Of an organism 
or organ which perfectly fulfilled, in its structure and work- 
ing, the end of its existence, nothing further required to be 
said or known. Physiologists of the present day recognise 
as fully as their predecessors that perfection of contrivance 
which displays itself in all living structures, the more ex- 
quisitely the more minutely they are examined. No one, 
for example, has written more emphatically on this point 
than did Ludwig. In one of his discourses, after showing 
how Nature exceeds the highest standard of human attain- 
ment — how she fashions as it were out of nothing and with- 
out tools, instruments of a perfection which the human 
artificer cannot reach, though provided with every suitable 
material — wood, brass, glass, india-rubber — he gives the 
organ of sight as a signal example, referring among its 

1 1 am aware that in thus stating the relation between adaptation and the 
struggle for existence, I may seem to be reversing the order followed by 
Mr. Darwin, inasmuch as he regarded the survival of organisms which are 
fittest for their place in Nature, and of parts which are fittest for their 
place in the organism, as the agency by which adaptedness is brought about. 
However this may be expressed it cannot be doubted that fitness is an 
essential of organisms. Living beings are the only things in Nature which 
by virtue of evolution and descent are able to adapt themselves to their 
surroundings. It is therefore only so far as organism (with all its attri- 
butes) is presupposed, that the dependence of adaptation on survival is 


other perfections to the rapidity with which the eye can 
be fixed on numerous objects in succession and the instan- 
taneous and unconscious estimates which we are able to 
form of the distances of objects, each estimate involving a 
process of arithmetic which no calculating machine could 
effect in the time. 1 In another discourse — that given at 
Leipzig when he entered on his professorship in 1865 — he re- 
marks that when in our researches into the finer mechanism 
of an organ we at last come to understand it, we are 
humbled by the recognition " that the human inventor is 
but a blunderer as compared with the unknown Master of 
the animal creation ", 2 

Some readers will perhaps remember how one of the 
most brilliant of philosophical writers, in a discourse to the 
British Association delivered a quarter of a century ago, 
averred on the authority of a great Physiologist that the 
eye, regarded as an optical instrument, was so inferior a 
production that if it were the work of a mechanician it 
would be unsaleable. Without criticising or endeavouring 
to explain this paradox, I may refer to it as having given 
the countenance of a distinguished name to a misconception 
which I know exists in the minds of many persons, to the 
effect that the scientific Physiologist is more or less blind to 
the evidence of design in creation. On the contrary, the 
view taken by Ludwig, as expressed in the words I have 
quoted, is that of all Physiologists. The disuse of the 
teleological expressions which were formerly current does 
not imply that the indications of contrivance are less ap- 
preciated, for, on the contrary, we regard them as more 
characteristic of organism as it presents itself to our obser- 
vation than any other of its endowments. But, if I may 

1 I summarise here from a very interesting lecture entitled " Leid und 
Freude in der Naturforschung " published in the Gartenlaube (Nos. 22 and 
23) in 1870. 

2 The words translated in the above sentence are as follows : " Wenn 
uns endlich die Palme gereicht wird, wenn wir ein Organ in seinem 
Zuzammenhang begreifen, so wird unser stolzes Gattungsbewusstsein durch 
die Erkenntniss niedergedruckt, dass der menschlicher Erfinder ein Stumper 
gegen den unbekannten Meister der thierischen Schbpfung sei ". 


be permitted to repeat what has been already said, we use 
the evidences of adaptation differently. We found no ex- 
planation on this or any other biological principle, but refer 
all the phenomena by which these manifest themselves to 
the simpler and more certain Physical Laws of the Universe. 

Why must we take this position ? First, because it is a 
general rule in investigations of all kinds to explain the 
more complex by the more simple. The material Universe 
is manifestly divided into two parts, the living and the non- 
living. We may, if we like, take the living as our Norma, 
and say to the Physicists, You must come to us for Laws, 
you must account for the play of energies in universal nature 
by referring them to Evolution, Descent, Adaptation. Or 
we may take these words as true expressions of the mutual 
relations between the phenomena and processes peculiar to 
living beings, using for the explanation of the processes 
themselves the same methods which we should employ if 
we were engaged in the investigation of analogous pro- 
cesses going on independently of life. Between these two 
courses there seems to me to be no third alternative, unless 
we suppose that there are two material Universes, one to 
which the material of our bodies belongs, the other com- 
prising everything that is not either plant or animal. 

The second reason is a practical one. We should have 
to go back to the time which I have ventured to call pre- 
scientific, when the world of life and organisation was sup- 
posed to be governed exclusively by its own Laws. The 
work of the past fifty years has been done on the opposite 
principle, and has brought light and clearness where there 
was before obscurity and confusion. All this progress we 
should have to repudiate, but this would not be all. We 
should have to forego the prospect of future advance. 
Whereas by holding on our present course, gradually pro- 
ceeding from the more simple to the more complex, from 
the physical to the vital, we may confidently look forward 
to extending our knowledge considerably beyond its present 

A no less brilliant writer than the one already referred 
to, who is also no longer with us, asserted that mind was a 


secretion of the brain in the same sense that bile is a secretion 
of the liver or urine that of the kidney ; and many people 
have imagined this to be the necessary outcome of a too 
mechanical way of looking at vital phenomena, and that 
Physiologists, by a habit of adhering strictly to their own 
method, have failed to see that the organism presents prob- 
lems to which this method is not applicable, such, e.g., as 
the origin of the organism itself, or the origin and develop- 
ment in it of the mental faculty. The answer to this sug- 
gestion is that these questions are approached by Physio- 
logists only in so far as they are approachable. We are 
well aware that our business is with the unknown knowable, 
not with the transcendental. During the last twenty years 
there has been a considerable forward movement in Physio- 
logy in the psychological direction, partly dependent on 
discoveries as to the localisation of the higher functions of 
the nervous system, partly on the application of methods of 
measurement to the concomitant phenomena of psychical 
processes. And these researches have brought us to the 
very edge of a region which cannot be explored by our 
methods — where measurements of time or of space are no 
longer possible. 

In approaching this limit the Physiologist is liable to fall 
into two mistakes — on the one hand, that of passing into 
the transcendental without knowing it ; on the other, that 
of assuming that what he does not know is not knowledge. 
The first of these risks seems to me of little moment ; first, 
because the limits of natural knowledge in the psychological 
direction have been well defined by the best writers, as, e.g., 
by du Bois-Reymond in his well-known essay "On the 
Limits of Natural Knowledge," but chiefly because the in- 
vestigator who knows what he is about is arrested in limine 
by the impossibility of applying the experimental method 
to questions beyond its scope. The other mistake is chiefly 
fallen into by careless thinkers, who, while they object to 
the employment of intuition even in regions where intuition 
is the only method by which anything can be learned, 
attempt to describe and define mental processes in mechan- 
ical terms, assigning to these terms meanings which science 


does not recognise, and thus slide into a kind of speculation 
which is as futile as it is unphilosophical. 


The uneventful history of Ludwig's life — how early he 
began his investigation of the anatomy and function of the 
kidneys ; how he became just fifty years ago titular Pro- 
fessor at Marburg, in the small University of his native 
State, Hesse Cassel ; how in 1849 he removed to Zurich as 
actual Professor and thereupon married ; how he was six 
years later promoted to Vienna — has already been admirably 
related in these pages by Dr. Stirling. In 1865, after 
twenty years of professorial experience, but still in the 
prime of life and, as it turned out, with thirty years of 
activity still before him, he accepted the Chair of Physio- 
logy at Leipzig. His invitation to that great University 
was by far the most important occurrence in his life, for the 
liberality of the Saxon Government, and particularly the 
energetic support which he received from the enlightened 
Minister v. Falkenstein, enabled him to accomplish for 
Physiology what had never before been attempted on an 
adequate scale. No sooner had he been appointed than 
he set himself to create what was essential to the 
progress of the Science — a great Observatory, arranged 
not as a Museum, but much more like a physical and 
chemical Laboratory, provided with all that was needed for 
the application of exact methods of research to the investiga- 
tion of the processes of Life. The idea which he had ever in 
view, and which he carried into effect during the last thirty 
years of his life with signal success, was to unite his life- 
work as an investigator with the highest kind of teaching. 
Even at Marburg and at Zurich he had begun to form a 
School ; for already men nearly of his own age had rallied 
round him. Attracted in the first instance by his early 
discoveries, they were held by the force of his character, 
and became permanently associated with him in his work 
as his loyal friends and followers — in the highest sense his 
scholars. If, therefore, we speak of Ludwig as one of the 


greatest teachers of Science the world has seen, we 
have in mind his relation to the men who ranged them- 
selves under his leadership in the building up of the Science 
of Physiology, without reference to his function as an 
ordinary academical teacher. 

Of this relation we can best judge by the careful perusal 
of the numerous biographical memoirs which have appeared 
since his death, more particularly those of Professor His 1 
(Leipzig), of Professor Kronecker 2 (Bern), who was for 
many years his coadjutor in the Institute, of Professor v. 
Fick 3 (Wlirzburg), of Professor v. Kries * (Freiburg), of 
Professor Mosso 5 (Turin), of Professor Fano 6 (Florence), 
of Professor Tigerstedt 7 (Upsala), of Professor Stirling s 
in England. With the exception of Fick, whose relations with 
Ludwig were of an earlier date, and of his colleague in the 
Chair of Anatomy, all of these distinguished teachers were 
at one time workers in the Leipzig Institute. All testify 
their love and veneration for the master, and each contributes 
some striking touches to the picture of his character. 

All Ludwig's investigations were carried out with his 
scholars. He possessed a wonderful faculty of setting each 
man to work at a problem suited to his talent and previous 
training, and this he carried into effect by associating him 
with himself in some research which he had either in 
progress or in view. During the early years of the Leip- 
zig period, all the work done under his direction was 
published in the well-known volumes of the Arbeiten, and 

1 His. " Karl Ludwig und Karl Thiersch.'' Akademische Geddcht- 
nissrede, Leipzig, 1S95. 

2 Kronecker. "Carl Friedrich Wilhelm Ludwig." Berliner Klin, 
Wochensch., 1895, No. 21. 

3 A. Fick. " Karl Ludwig." Nachruf. Biographische Blatter, Berlin, 
vol. i., pt. 3. 

4 v. Kries. "Carl Ludwig." Freiburg, Bd. i., 1895. 

5 Mosso. " Karl Ludwig." Die Nation, Berlin, Nos. 38, 39. 

6 Fano. " Per Carlo Ludwig Commemorazione." Clinica Afodema, 
Florence, i., No. 7. 

7 Tigerstedt. " Karl Ludwig." Denkrede. Biographische Blatter, Berlin, 
vol. i., pt. 3. 

8 Stirling. "Science Progress," vol. iv., No. 21. 


subsequently in the Archiv fur Anat. unci Physiologic of 
du Bois-Reymoncl. Each " Arbeit ' : of the laboratory 
appeared in print under the name of the scholar who 
operated with his master in its production, but the 
scholar's part in the work done varied according to its 
nature and his ability. Sometimes, as v. Kries says, he sat 
on the window-sill while Ludwig with the efficient help of 
his laboratory assistant Salvenmoser, did the whole of the 
work. In all cases Ludwig not only formulated the 
problem, but indicated the course to be followed in each 
step of the investigation, calling the worker, of course, into 
counsel. In the final working up of the results he always 
took a principal part, and often wrote the whole paper. But 
whether he did little or much, he handed over the whole 
credit of the performance to his coadjutor. This method of 
publication has no doubt the disadvantage that it leaves 
it uncertain what part each had taken ; but it is to be 
remembered that this drawback is unavoidable whenever 
master and scholar work together, and is outweighed by the 
many advantages which arise from this mode of co-opera- 
tion. The instances in which any uncertainty can exist in 
relation to the real authorship of the Leipzig work are 
exceptional. The well-informed reader does not need to 
be told that Mosso or Schmidt, Brunton or Gaskell, Stirling 
or Wooldridge were the authors of their papers in a sense 
very different from that in which the term could be applied 
to some others of Ludwig's pupils. On the whole the plan 
must be judged of by the results. It was by working with 
his scholars that Ludwig trained them to work afterwards by 
themselves ; and thereby accomplished so much more than 
other great teachers have done. 

I do not think that any of Ludwig's contemporaries 
could be compared to him in respect of the wide range of 
his researches. In a science distinguished from others by 
the variety of its aims, he was equally at home in all 
branches, and was equally master of all methods, for he 
recognised that the most profound biological question can 
only be solved by combining anatomical, physical and 
and chemical inquiries. It was this consideration which led 


him in planning the Leipzig Institute to divide it into three 
parts, experimental (in the more restricted sense), chemical 
and histological. Well aware that it was impossible for a 
man who is otherwise occupied to maintain his familiarity 
with the technical details of Histology and Physiological 
Chemistry, he placed these departments under the charge 
of younger men capable of keeping them up to the rapidly 
advancing standard of the time, his relations with his 
coadjutors being such that he had no difficulty in retaining 
his hold of the threads of the investigation to which these 
special lines of inquiry were contributory. 

It is scarcely necessary to say that as an experimenter 
Ludwig was unapproachable. The skill with which he 
carried out difficult and complicated operations, the care 
with which he worked, his quickness of eye and certainty 
of hand were qualities which he had in common with great 
surgeons. In employing animals for experiment he strongly 
objected to rough and ready methods, comparing them to 
" firing a pistol into a clock to see how it works ". Every 
experiment ought, he said, to be carefully planned and 
meditated on beforehand, so as to accomplish its scientific 
purpose and avoid the infliction of pain. To ensure this 
he performed all operations himself, only rarely committing 
the work to a skilled coadjutor. 

His skill in anatomical work was equally remarkable. 
It had been acquired in early days, and appeared throughout 
his life to have given him very great pleasure, for Mosso 
tells how, when occupying the room adjoining that in which 
Ludwig was working as he usually did by himself, he heard 
the outbursts of glee which accompanied each successful step 
in some difficult anatomical investigation. 

Let us now examine more fully the part which Ludwig 
played in the revolution of ideas as to the nature of vital 
processes which, as we have seen, took place in the middle 
of the present century. 

Although, as we shall see afterwards, there were many 
men who, before Ludwig's time, investigated the phenomena 
of life from the physical side, it was he and the contem- 
poraries who were associated with him who first clearly 


recognised the importance of the principle that vital pheno- 
mena can only be understood by comparison with their physical 
counterparts, and foresaw that in this principle the future of 
Physiology was contained as in a nutshell. Feeling strongly 
the fruitlessness and unscientific character of the doctrines 
which were then current, they were eager to discover 
chemical and physical relations in the processes of life. 
In Ludwig's intellectual character this eagerness expressed 
his dominant motive. Notwithstanding that his own re- 
searches had in many instances proved that there are im- 
portant functions and processes in the animal organism 
which have no physical or chemical analogues, he never 
swerved either from the principle or from the method 
founded upon it. 

Although Ludwig was strongly influenced by the rapid 
progress which was being made in scientific discovery at 
the time that he entered on his career, he derived little from 
his immediate predecessors in his own science. He is 
sometimes placed among the pupils of the great Comparative 
Anatomist and Physiologist, J. Mliller. This, however, is a 
manifest mistake, for Ludwig did not visit Berlin until 1847, 
when Miiller was nearly at the end of his career. At that 
time he had already published researches of the highest 
value (those on the Mechanism of the Circulation and on the 
Physiology of the Kidney), and had set forth the line in 
which he intended to direct his investigations. The only 
earlier Physiologist with whose work that of Ludwig can 
be said to be in real continuity was E. H. Weber, whom he 
succeeded at Leipzig, and strikingly resembled in his way 
of working. For Weber, Ludwig expressed his veneration 
more unreservedly than for any other man, excepting per- 
haps Helmholtz, regarding his researches as the foundation 
on which he himself desired to build. Of his colleagues at 
Marburg he was indebted in the first place to the anatomist, 
Professor Ludwig Fick, in whose department he began his 
career as Prosector, and to whom he owed facilities without 
which he could not have carried out his earlier researches ; 
and in an even higher degree to the great chemist, R. W. 
Bunsen, from whom he derived that training in the exact 


sciences which was to be of such inestimable value to him 

There is reason, however, to believe that, as so often 
happens, Ludwig's scientific progress was much more in- 
fluenced by his contemporaries than by his seniors. In 
1847, as we learn on the one hand from du Bois-Reymond, 
on the other from Ludwig himself, he visited Berlin for the 
first time. This visit was an important one both for him- 
self and for the future of Science, for he there met three 
men of his own age, Helmholtz, du Bois-Reymond and 
Brticke, who were destined to become his life-friends, all of 
whom lived nearly as long as Ludwig himself, and attained 
to the highest distinction. They all were full of the same 
enthusiasm. As Ludwig said when speaking of this visit : 
" We four imagined that we should constitute Physiology 
on a chemico-physical foundation, and give it equal scientific 
rank with Physics, but the task turned out to be much more 
difficult than we anticipated ". These three young men, 
who were devoted disciples of the great Anatomist, had the 
advantage over their master in the better insight which 
their training had given them into the fundamental prin- 
ciples of scientific research. They had already gathered 
around themselves a so-called "physical " school of Physio- 
logy, and welcomed Ludwig on his arrival from Marburg 
as one who had of his own initiative undertaken in his own 
University das Befremngswerk aus dcm Vitalismus. 

The determination to refer all vital phenomena to their 
physical or chemical counterparts or analogues, which, as I 
have said, was the dominant motive in Ludwio's char- 
acter, was combined with another quality of mind which if 
not equally influential was even more obviously displayed in 
his mode of thinking and working. His first aim, even 
before he sought for any explanation of a structure or of 
a process, was to possess himself, by all means of observa- 
tion at his disposal, of a complete objective conception of 
all its relations. He regarded the faculty of vivid sensual 
realisation (lebendige sinnliche Anschanung) as of special 
value to the investigator of natural phenomena, and did his 
best to cultivate it in those who worked with him in the 


laboratory. In himself, this objective tendency (if I may 
be permitted the use of a word which, if not correct, seems 
to express what I mean) might be regarded as almost a 
defect, for it made him indisposed to appreciate any sort of 
knowledge which deals with the abstract. He had a 
disinclination to philosophical speculation which almost 
amounted to aversion, and, perhaps for a similar reason, 
avoided the use of mathematical methods even in the 
discussion of scientific questions which admitted of being 
treated mathematically — contrasting in this respect with 
his friend du Bois-Reymond, resembling Brlicke. But 
as a teacher the quality was of immense use to him. His 
power of vivid realisation was the substratum of that many- 
sidedness which made him, irrespectively of his scientific 
attainments, so attractive a personality. 

I am not sure that it can be generally stated that a keen 
scientific observer is able to appreciate the artistic aspects 
of Nature. In Ludwig's case, however, there is reason to 
think that aesthetic faculty was as developed as the power of 
scientific insight. He was a skilful draughtsman but not 
a musician ; both arts were, however, a source of enjoy- 
ment to him. He was a regular frequenter of the Gewand- 
kaus concerts, and it was his greatest pleasure to bring to- 
gether gifted musicians in his house, where he played the 
part of an intelligent and appreciative listener. Of painting 
he knew more than of music, and was a connoisseur whose 
opinion carried weight. It is related that he was so worried 
by what he considered bad art, that after the redecoration 
of the Gczvandhaus concert-room, he was for some time 
deprived of his accustomed pleasure in listening to music. 

Ludwig's social characteristics can only be touched on 
here in so far as they serve to make intelligible his wonder- 
ful influence as a teacher. Many of his pupils at Leipzig 
have referred to the schbne Gemeinsamkeit which char- 
acterised the life there. The harmonious relation which, 
as a rule, subsisted between men of different education and 
different nationalities, could not have been maintained had 
not Ludwig possessed side by side with that inflexible 
earnestness which he showed in all matters of work or 


duty a certain youthfulness of disposition which made it 
possible for men much younger than himself to accept his 
friendship. This sympathetic geniality was, however, not 
the only or even the chief reason why Ludwig's pupils were 
the better for having known him. There were not a few 
of them who for the first time in their lives came into 
personal relation with a man who was utterly free from 
selfish aims and vain ambitions, who was scrupulously 
conscientious in all that he said and did, who was what he 
seemed, and seemed what he was, and who had no other 
aim than the advancement of his science, and in that ad- 
vancement saw no other end than the increase of human 
happiness. These qualities displayed themselves in Lud- 
wig's daily active life in the laboratory, where he was to be 
found whenever work of special interest was going on ; but 
still more when, as happened on Sunday mornings, he was 
"at home" in the library of the Institute — the corner room 
in which he ordinarily worked. Many of his "scholars" 
have put on record their recollections of these occasions, the 
cordiality of the master's welcome, the wide range and 
varied interest of his conversation, and the ready apprecia- 
tion with which he seized on anything that was new or 
original in the suggestions of those present. Few men 
live as he did, " im Gaznen, Gtiten, Sckonen" and of those 
still fewer know how to communicate out of their fulness to 


Since the middle of the century the progress of Physio- 
logy has been continuous. Each year has had its record, 
and has brought with it new accessions to knowledge. In 
one respect the rate of progress was more rapid at first than 
it is now, for in an unexplored country discovery is relatively 
easy. In another sense it was slower, for there are now 
scores of investigators for every one that could be counted 
in 1840 or 1850. Until recently there has been throughout 
this period no tendency to revert to the old methods — no 
new departure — no divergence from the principles which 
Ludwig did so much to enforce and exemplify. 


The wonderful revolution which the appearance of the 
Origin of Species produced in the other branch of Biology, 
promoted the progress of Physiology, by the new interest 
which it gave to the study, not only of structure and de- 
velopment, but of all other vital phenomena. It did not, 
however, in any sensible degree affect our method or alter 
the direction in which Physiologists had been working for 
two decades. Its most obvious effect was to sever the two 
subjects from each other. To the Darwinian epoch Com- 
parative Anatomy and Physiology were united, but as the 
new Ontology grew, it became evident that each had its own 
problems and its own methods of dealing with them. 

The old vitalism of the first half of the century is easily 
explained. It was generally believed that, on the whole, 
things went on in the living body as they do outside of 
it, but when a difficulty arose in so explaining them the 
Physiologist was ready at once to call in the aid of a 
" vital force' '. It must not, however, be forgotten that, as I 
have already indicated, there were great teachers (such, 
for example, as Sharpey and Allen Thomson in England, 
Magendie in France, Weber in Germany) who discarded 
all vitalistic theories, and concerned themselves only with 
the study of the time- and place-relations of phenomena ; 
men who were before their time in insight, and were only 
hindered in their application of chemical and physical prin- 
ciples to the interpretation of the processes of life by the 
circumstance that chemical and physical knowledge was in 
itself too little advanced. Comparison was impossible, for 
the standards were not forthcoming. 

Vitalism in its original form gave way to the rapid ad- 
vance of knowledge as to the correlation of the physical 
sciences which took place in the forties. Of the. many 
writers and thinkers who contributed to that result, J. R. 
Mayer and Helmholtz did so most directly, for the con- 
tribution of the former to the establishment of the Doctrine 
of the Conservation of Energy had physiological considera- 
tions for its point of departure ; and Helmholtz, at the time 
he wrote the Erhaltung der Kraft, was still a Physiolo- 
gist. Consequently when Ludwig's celebrated Lehrbuch 


came out in 1852, the book which gave the coup de grace to 
vitalism in the old sense of the word, his method of setting 
forth the relations of vital phenomena by comparison with 
their physical or chemical counterparts, and his assertion that 
it was the task of Physiology to make out their necessary 
dependence on elementary conditions, although in violent 
contrast with current doctrine, were in no way surprising to 
those who were acquainted with the then recent progress 
of research. Ludwig's teaching was indeed no more than 
a general application of principles which had already been 
applied in particular instances. 

The proof of the non-existence of a special " vital force " 
lies in the demonstration of the adequacy of the known 
sources of energy in the organism to account for the actual 
day by day expenditure of heat and work — in other words, 
on the possibility of setting forth an energy balance sheet in 
which the quantity of food which enters the body in a given 
period (hour or day) is balanced by an exactly correspond- 
ing amount of heat produced or external work done. It is 
interesting to remember that the work necessary for 
preparing such a balance sheet (which Mayer had attempted, 
but, from want of sufficient data, failed in) was begun 
thirty years ago in the laboratory of the Royal Institution 
by the Foreign Secretary of the Royal Society. But the 
determinations made by Dr. Frankland related to one side of 
the balance sheet, that of income. By his researches in 1 866 
he gave Physiologists for the first time reliable information 
as to the heat value {i.e., the amount of heat yielded by the 
combustion) of different constituents of food. It still re- 
mained to apply methods of exact measurement to the 
expenditure side of the account. Helmholtz had estimated 
this, as regards man, as best he might, but the technical 
difficulties of measuring the expenditure of heat of the 
animal body appeared until lately to be almost insuperable. 
Now that it has been at last successfully accomplished, we 
have the experimental proof that in the process of life there 
is no production or disappearance of energy. It may be 
said that it was unnecessary to prove what no scientifically 

sane man doubted. There are, however, reasons why it is 



of importance to have objective evidence that food is the 
sole and adequate source of the energy which we day by 
day or hour by hour disengage, whether in the form of heat 
or external work. 

In the opening paragraph of this section it was observed 
that until recently there had been no tendency to revive the 
vitalistic notion of two generations ago. In introducing the 
words in italics I referred to the existence at the present 
time in Germany of a sort of reaction, which under the 
term " Neovitalismus " has attracted some attention — -so 
much indeed that at the Versani7nlung Deutscher Natur- 
forscher at Ltibeck last September, it was the subject of 
one of the general addresses. The author of this address, 
Prof. Rindfleisch, was, I believe, the inventor of the word ; 
but the origin of the movement is usually traced to a work 
on Physiological Chemistry which an excellent translation 
by the late Dr. Wooldridge has made familiar to English 
students. The author of this work owes it to the language 
he employs in the introduction on " Mechanism and 
Vitalism," if his position has been misunderstood, for in 
that introduction he distinctly ranges himself on the vital- 
istic side. As, however, his vitalism is of such a kind as 
not to influence his method of dealing with actual problems, 
it is only in so far of consequence as it may affect the reader. 
For my own part I feel grateful to Professor Bange for 
having produced an interesting and readable book on a dry 
subject, even though that interest may be partly due to the 
introduction into the discussion of a question which, as he 
presents it, is more speculative than scientific. 

As regards other physiological writers to whom vitalistic 
tendencies have been attributed, it is to be observed that 
none of them have even suggested that the doctrine of a 
"vital force" in its old sense should be revived. Their 
contention amounts to little more than this, that in certain 
recent instances improved methods of research appear to 
have shown that processes at first regarded as entirely 
physical or chemical do not conform so precisely as they 
were expected to do to chemical and physical laws. As 
these instances are all essentially analogous, reference to 
one will serve to explain the bearing of the rest. 


Those who have any acquaintance with the structure of 
the animal body will know that there exists in the higher 
animals, in addition to the system of veins by which the 
blood is brought back from all parts to the heart, another 
less considerable system of branched tubes, the lymphatics, 
by which, if one may so express it, the leakage of the blood- 
vessels is collected. Now, without inquiring into the why 
of this system, Ludwig and his pupils made and continued 
for many years elaborate investigations which were for long 
the chief sources of our knowledge, their general result 
being that the efficient cause of the movement of the lymph, 
like that of the blood, was mechanical. At the Berlin Con- 
gress in 1890 new observations by Professor Heidenhain of 
Breslau made it appear that under certain conditions the 
process of lymph formation does not go on in strict accord- 
ance with the physical laws by which leakage through 
membranes is regulated, the experimental results being of 
so unequivocal a kind that, even had they not been con- 
firmed, they must have been received without hesitation. 
How is such a case as this to be met? The "Neovitalists " 
answer promptly by reminding us that there are cells, i.e., 
living individuals, placed at the inlets of the system of 
drainage without which it would not work, that these let in 
less or more liquid according to circumstances, and that in 
doing so they act in obedience, not to physical laws, but to 
vital ones — to internal laws which are special to themselves. 

Now, it is perfectly true that living cells, like working 
bees, are both the architects of the hive and the sources of 
its activity, but if we ask how honey is made it is no answer 
to say that the bees make it. We do not require to be told 
that cells have to do with the making of lymph as with 
every process in the animal organism, but what we want to 
know is how they work, and to this we shall never get an 
answer so long as we content ourselves with merely ex- 
plaining one unknown thing by another. The action of 
cells must be explained, if at all, by the same method of 
comparison with physical or chemical analogues that we 
employ in the investigation of organs. 

Since 1890 the problem of lymph formation has been 


attacked by a number of able workers, among others here 
in London, by Dr. Starling of Guy's Hospital, who, by 
sedulously studying the conditions under which the dis- 
crepancies between the actual and the expected have arisen, 
has succeeded in untying several knots. In reference to 
the whole subject, it is to be noticed that the process by 
which difficulties are brought into view is the same as that 
by which they are eliminated. It is one and the same 
method throughout, by which step by step, knowledge per- 
fects itself — at one time by discovering errors, at another 
by correcting them ; and if at certain stages in this pro- 
gress difficulties seem insuperable, we can gain nothing by 
calling in, even provisionally, the aid of any sort of Eidolon, 
whether "cell," "protoplasm" or internal principle. 

It thus appears to be doubtful whether any of the 
biological writers who have recently professed vitalistic 
tendencies are in reality vitalists. The only exception 
that I know is to be found in the writings of a well- 
known morphologist, Dr. Hans Driesch, 1 who has been 
led by his researches on what is now called the Me- 
chanics of Evolution to revert to the fundamental con- 
ception of vitalism, that the laws which govern vital 
processes are not physical, but biological — that is, peculiar 
to the living organism, and limited thereto in their 
operation. Dr. Driesch's researches as to the modifi- 
cations which can be produced by mechanical inter- 
ference in the early stages of the process of ontogenesis 
have enforced upon him considerations which he evidently 
regards as new, though they are familiar enough to Physio- 
logists. He recognises that although by the observation of 
the successive stages in the ontogenetic process, one may 
arrive at a perfect knowledge of the relation of these stages 
to each other, this leaves the efficient causes of the develop- 
ment unexplained [fukrt nicht zu einem Erkenntniss ihrer 
bewirkenden Ursacheii)- — it does not teach us why one 

1 Driesch. " Entwicklungsmechanische Studien " : a series of ten 
Papers, of which the first six appeared in the Zeitsch. /. w. Zoologie, vols, 
liii. and lv. ; the rest in the Mittheihingen of the Naples Station. 


form springs out of another. This brings him at once face 
to face with a momentous question. He has to encounter 
three possibilities — he may either join the camp of the 
biological agnostics and say with du Bois-Reymond, "ignora- 
mus et ignorabimus" or be content to work on in the hope 
that the physical laws that underlie and explain organic 
Evolution may sooner or later be discovered, or he may 
seek for some hitherto hidden Law of Organism of which 
the known facts of Ontogenesis are the expression, and 
which, if accepted as a Law of Nature, would explain every- 
thing. Of the three alternatives Driesch prefers the last, 
which is equivalent to declaring himself an out and out 
vitalist. He trusts by means of his experimental investiga- 
tions of the Mechanics of Evolution to arrive at " elementary 
conceptions" on which by "mathematical deduction" 1 a 
complete theory of Evolution may be founded. 

If this anticipation could be realised, if we could con- 
struct with the aid of those new Principia the ontogeny of 
a single living being, the question whether such a result 
was or was not inconsistent with the uniformity of Nature, 
would sink into insignificance as compared with the 
splendour of such a discovery. 

But will such a discovery ever be made ? It seems to 
me even more improbable than that of a physical theory 
of organic evolution. It is satisfactory to reflect that the 
opinion we may be led to entertain on this theoretical 
question need not affect our estimate of the value ol Dr. 
Driesch's fruitful experimental researches. 

J. Burdon Sanderson. 

1 " Elementarvorstellungen . . . die zwar mathematische Deduktion 
aller Erscheinungen aus sich gestatten mochten." Driesch. " Beitrage 
zur theoretischen Morphologic" Biol. Centralblatt, vol. xii., p. 539, 1892. 




DURING the last quarter of a century a considerable 
change has passed over the aspect of biology, 
especially in this country. It was formerly possible for a 
man to be, fairly at any rate, well up in the two branches 
of zoology and botany, but this is no longer possible, 
regarded from our modern standpoint. Specialisation, 
inevitable owing to the rapid advances which have been 
everywhere made, has not only effected a practical 
divorce between these two sciences, but the same disrupting 
agency is operating continuously in each of them. 

None the less is it true, however, that there are certain 
features of fundamental importance which are shared alike 
by animals and plants. This community of structure is 
most clearly recognised within the limits of the individual 
cells, and it is perhaps nowhere more impressively demon- 
strated than in the remarkable similarity which exists 
between the nuclear division as observed in animals and in 
plants, — a similarity which may extend to the most minute 

The cell, using the word in its widest sense, is, as 
Haeckel said long ago, emphatically the unit of life. For 
though the several parts, such as nucleus and the cell- 
protoplasm, which together constitute a cell, all possess 
autonomy to a certain degree, it still remains true that it is 
only when they operate jointly and in harmony that a suc- 
cessful and "going concern," a living individual, is the 
result. And since we have strong reasons for believing 
that animals and plants represent the diverging limbs of a 
stock traceable at the root to a common source, viz., lowly 
unicellular organisms, it is obvious that the study of the cell, 
of its structure and of the functions discharged by its 
various parts, offers an immensely important, though it 
may well be a very difficult, field for research. 


What, we may ask, is the essential structure of the 
protoplasm, of the nucleus, and of those marvellous bodies, 
the chromosomes, which reappear at every nuclear division ? 
What is it that initiates the division of a cell or of its 
nucleus, and why do some cells go through such complex 
evolutions whilst others seem to adopt a relatively simple 
course? What is it that determines that the descendants of 
one cell shall develop differently from those of another, so 
as to give rise to this or that tissue system ? Or again, 
how is the unicellular condition of an infusorian compatible 
with an intricate and often highly differentiated organisa- 
tion ? 

These and a host of other questions rise and confront 
us on the very threshold of our inquiry, and the hints which 
Nature has dropped for our guidance are at best only 
obscure ones ; thus the position of the biological investigator 
contrasts unfavourably with that of the chemist or physicist, 
inasmuch as he is generally debarred, owing to the very 
conditions of the bodies he is dealing with, from having 
recourse to direct experiment ; Nature conducts the experi- 
ments and he has to remain content with watching- the 
result, analysing the factors and reconstructing the process 
as best he can. Nevertheless there is, clearly, no funda- 
mental distinction between the (so-called) observational 
and experimental sciences. 

It is, then, only by patient accumulation and careful 
comparison of all the facts that even a proximate solution 
of the difficulties before us can ever be reached. Much 
has been done in collecting the data, and a good deal is 
known both as to the structure of the cell and the phases 
through which it passes during its existence. And fortu- 
nately one generalisation is gradually emerging with in- 
creasing clearness from beneath the ever-growing pile of 
detail, and it promises to prove a guide of no small value, 
namely, that in those processes which we have reason to 
regard as fundamentally important there exists a surprising 
degree of similarity between the structural elements of 
animals on the one hand and of plants on the other. And 
these points of similarity are now known to be so numerous 


and so close that we are almost warranted in drawing the 
conclusion that the measure of the resemblance will afford a 
criterion as to the relative degree of importance to be 
attached to this or that phenomenon of cell life. 

It seems almost certain that this similarity is to be 
interpreted as the result of the evolution along parallel 
lines of a particular structural arrangement, or, to put it in 
another way, as being the outcome of the continuous opera- 
tion of similar forces upon an essentially similar proto- 
plasmic structure. No doubt all the change manifested in 
protoplasm is ultimately to be ascribed to the effects of 
forces upon its own material substance ; the special point 
of interest here lies in the similarity of the results. It 
cannot be due to mere accident that the stages in the develop- 
ment of the spermatozoa of a newt should bear a closer 
resemblance to the corresponding divisions in the pollen- 
mother-cell of a lily than they do to the rest of the tissue 
cells in the body of the same newt. 

In the present article it is not my purpose to attempt to 
summarise the vast amount of detail which has accumulated 
within recent years ; my aim is rather to try to indicate the 
general directions in which the results seem to be tending, 
and to point out the kind of evidence on which the current 
views are based. And although I am here especially dealing 
with the botanical aspect of the questions involved, it will 
be clear from what has been already said that it will be 
impossible, and certainly not desirable, to ignore the in- 
vestigations which have been prosecuted by the zoologists. 

And in order to make clear that which is to follow, it 
may not be superfluous to recapitulate the general relations 
of nucleus and cell protoplasm as commonly received at the 
present time. The essential character of all cells, whether 
animal or vegetable, and whether they exist as free inde- 
pendent organisms, or whether they form more or less 
highly differentiated colonies, consists in this, the association 
of a nucleus with a certain amount of cell protoplasm (com- 
monly called Cytoplasm, to distinguish it from the nuclear 
protoplasm). And this is equally true, so far as we have 
means of determining the question, in the case of those 


organisms in which we as yet have failed to recognise a de- 
finite nuclear body, for there are reasons for believing that the 
nuclear substance is in all cases really present, whether it 
happens to be collected into a specialised mass or not. And 
it should be remembered that the number of cells supposed 
to possess what we may term a distributed or discrete 
nucleus is becoming smaller as our means of investigations 
improve. Thus according to Wager ( 1 ) even Bacteria 
possess a true nucleus. 

I am perfectly aware that attacks have recently been 
made on the cell-theory as extended to explain the organisa- 
tion (Whitman, Sedgwick) of animals, and that nobody 
would assert the cell to the ultimate unit of living substance. 
But neither of these propositions really affects, or is con- 
cerned with, the point of view just now before us. We are 
not here dealing with the wide questions connected with the 
architecture of the organism as a whole, nor with the 
equally difficult one, as to what constitutes the ultimate 
units of living matter, rather we are content just now to 
study the interaction of the parts which together are capable 
of carrying on a continuous living existence, which form a 
living individual, and these parts consist jointly of the 
nucleus and its surrounding cytoplasm. 1 The occurrence of 
cell walls is a matter of no importance from a general stand- 
point, although when present they may profoundly modify 
the characters of the organism in which they are formed. 
Many plants are known in which the protoplasm is only 
delimited by a cell wall from the surrounding medium, while 
the oftentimes huge protoplasmic mass suffers no internal 
partitioning, although it contains a vast number of nuclei 
distributed through it. 

Sachs, with characteristic insight, long ago perceived 
that the presence or absence of cell walls is a matter of 
only secondary importance. Their sequence and arrange- 
ment at the time of their first appearance can be predicted 

1 The researches of Klebs, Acqua, and others have shown that although 
protoplasm deprived of a nucleus may sometimes even assimilate food and 
maintain life for a not inconsiderable period of time, it is incapable of 


from simple geometrical considerations quite independently 
of the ultimate form which will be finally assumed as 
the result of specialised growth. And in applying the word 
Non-cellular to those plants in which partition walls do not 
occur, he merely gives formal expression to the fact that 
these anatomical structures are absent, although in other 
respects the plants in question conform with those usually 
called multicellular, and they are not at all to be regarded as 
consisting of a single enlarged cell. In fact he has expressly 
stated that non-cellular plants are really the equivalent of 
multicellular organisms in which the formation of internal 
cell walls does not occur. More recently he has introduced 
the term Energid (2) to express the physiological individu- 
ality of those units I have here continued to call cells, and 
he thereby emphasises the fact of their real existence 
whether any positive anatomical boundaries can be dis- 
cerned between them or not. 

It must however be clearly understood that in formulat- 
ing the expression energid, Sachs lays especial stress on 
the dynamical aspect of the relations existing between the 
cytoplasm and the nucleus. But it will be admitted by most 
people that a conception of force apart from the material 
substance on or through which it acts, and by which its 
operation becomes perceptible to the senses, belongs to the 
domain of purely abstract ideas. We require to know far 
more of the nature and structure of protoplasm before we 
can usefully divorce our conceptions of force from our ex- 
perience of matter in attempting to ascertain the nature 
of those physiological causes of which all external form is 
but the outward and visible sign. Sachs himself, however, 
escapes the charge of vagueness, by restricting the applica- 
tion of his expression so as to impose a territorial limit to 
the sphere of influence mutually existing between each 
nucleus and the surrounding cytoplasm. For him the word 
Energid embodies the idea that the whole protoplasmic 
region is partitioned into smaller provinces each dominated 
by its own nucleus. And although it may be advantageous 
for the seprovinces to be delimited from each other by cell 
walls, permitting thereby a more complete independence to 


attach to each one severally, the existence of such well- 
defined boundaries is by no means an indispensable condition 
of great complexity of organisation. Caulerpa amongst the 
algae imitates very closely the differentiated form of some 
of the higher terrestrial plants, without however possessing 
their corresponding internal structure. Its protoplasm is 
bounded by an external wall only, and is not internally 
partitioned. And yet the characters distinctive of the 
energids in the leaf-like parts are assuredly different from 
those of the energids which exist in the creeping stem or 
rootlike fibres. A transition from the condition of Caulerpa 
to that of the higher plants may be seen in Cladophora, in 
which the filamentous body seems, at first sight, to be made 
up of chains of cells, each of which stands in a definite 
relation to the general symmetry of the branched plant ; 
nevertheless, closer examination shows that each "cell" is 
multi-nucleate, and really represents a federation of energids 
which so act together as to constitute morphological units as 
far as the external form of the plant as a whole is concerned. 
Sachs' conception of the energid has been assailed by 
some writers, and he has to some extent perhaps invited 
criticism by formerly affixing a quasi-morphological, as well 
as a physiological significance to the term. At first sight 
it may seem difficult to justify its application in those cases 
in which streaming movement happens to go on in certain 
layers of the protoplasm, whilst the layer in which the 
nuclei are embedded is at rest. It is obvious that if we 
admit, as we can hardly avoid doing, that the nucleus does 
really exert a directive action over a localised area, the 
migratory protoplasm (assuming the movement to affect the 
protoplasm, and not merely the granular bodies contained 
in it) must be constantly coming within the range of 
fresh centres of influence. It may perhaps be compared 
to the case of a person passing from a region presided 
over by one government into one under the jurisdic- 
tion of another. Such a person would naturally be subjected 
to changed conditions, without however affecting either his 
own identity or that of the particular political centres through 
which he may happen to travel. 


Strasburger (3) has attempted to define more clearly the 
position of the individual energid, by proposing to limit its 
application to the nucleus together with a special part of the 
cytoplasm which he calls Kinoplasm and which he regards 
as the proximate seat of the effective manifestation of the 
forces at work in the cell. He regards the nomadic 
streaming protoplasm as being mainly charged with the 
function of providing nourishment for the nucleus and 
kinoplasm, and he distinguishes it by the special term of 
Trophoplasm. Strasburger maintains this same distinction 
between the active Kinoplasm and the nutritive tropho- 
plasm in those cases in which the limits of the several 
energids correspond with those of the individual cells ; 
and in this he is logical enough, for we know that living 
cells are not isolated from each other, but that protoplasmic 
continuity exists between adjacent cells by means of pores 
in the intervening walls. How far the distinction between 
kinoplasm and trophoplasm is either justified by observa- 
tion or demanded by theory is another matter altogether. 

But although the conception of energids is a happy one, 
as enabling us to distinguish discrete individualities in what 
may at first sight appear to consist of a common structure, 
it is not to be inferred that the individuals enjoy independ- 
ence. The great merit of the idea lies in the fact that it 
serves to narrow down, and hence to render more clearly 
comprehensible, many important problems which call for a 
solution before we can hope to grapple successfully with 
the more advanced questions relating to those forces of a 
still higher order which control and apparently direct the 
development of the organism as a whole, or to put it in 
another way, which determine the course of development 
which the particular energids shall follow. Such control is 
plainly apparent at every stage in the life of an organism. 
Why does growth take place symmetrically so that the 
energids, cells, or whatever we may choose to call them, so 
act in unison as to produce a " body fitly joined together 
and compacted by that which every joint supplieth, accord- 
ing to the effectual working in the measure of every 
part " ? Without some such assumption how is it 


possible to account for the fact that in certain embryos 
which have been mutilated, the surviving cells are enabled 
to so modify the course of their normal development as 
to make good the loss, and thus to form a perfect, if 
somewhat miniature organism ? For had there been no 
mutilation the cells thus concerned would unquestion- 
ably not have developed in the same way, but would have 
fulfilled the allotted task of merely providing for the genesis 
of their normal tissue products. Or again, why is it that 
when a lizard's tail is broken off the general form of the 
entire animal is once more reproduced, even though there 
are important histological and structural (but probably not 
functional) differences in the new tail as compared with 
that of the original one (4) ? 

When differentiation has so far become manifested in 
an organism that the limits of the several energids are 
coterminous with the cell walls, a considerable increase in 
their degree of independence doubtless ensues, but it is, as 
already stated, by no means absolute, and the examples just 
quoted support the statement. Whether organisation is 
the result of, or the factor which determines, the co-ordi- 
nate action of the cells is a question which we may safely 
leave to the future to decide. But perhaps it may be 
permissible to compare the cell colony which forms the 
organism to an isolated society in which the caste system 
prevails. Each caste or cell group is predestined to dis- 
charge certain definite offices in the state or the organism. 
If some indispensable caste should become exterminated, it is 
obvious that a differentiation and displacement must occur 
amongst those which survive, and this differentiation 
might either be readily complete, or it might only arise as 
a reluctant concession to necessity, just as a willow twig 
planted upside down in damp soil will form roots at this, its 
upper, end ; though comparison with a twig planted with 
its basal end in the ground will show how severe a tax the 
unusual effort has proved. 

It has already been said that an energid, and it might 
also be added, a typical cell, consists essentially of a nucleus 
and the protoplasm included within a certain area around 


it. But we cannot as yet answer the more obvious and, one 
might think, almost preliminary question as to what the 
chief functions which are discharged by these two com- 
ponents really may be. It is certain that the existence of a 
nucleus is essential to morphological development such as is 
implied in the production of new cells, and very probably 
also in the further differentiation of those which have 
already been formed. Instances of this are seen for example 
in the growth or alteration of the cell wall. Haberlandt 
(5) some years ago drew special attention to the fact 
that when local thickening occurred in a cell wall the 
nucleus commonly moved to this spot, and the present 
writer has repeatedly observed it during the formation of 
the hard coat found on many seeds ; here the deposition of 
substance is usually localised on the inner parts of the cell, 
and the nucleus takes up a corresponding position as soon 
as the process begins. Korschelt (6) has observed a 
similar relation to exist during the chitinisation of the mem- 
branes of insect cells, and quite recently Istvanffi (Ber. Deut. 
Gesel., Dec, 1895) has observed that when the tubular 
hypha of Mucor branches, a nucleus is invariably present at 
the spot whence the branch is arising. Strasqurger (3^) has 
also drawn attention to the same truth, inasmuch as he 
states that before the opening of the zoosporangium of 
CEdogonium, the nucleus and kinoplasm aggregate in the 
vicinity of the spot at which the hole is about to be formed. 

But perhaps one of the most striking instances of the 
directive effect of the nucleus as a whole is to be seen in 
the result of an experiment of Boveri, who asserts that he 
impregnated a non-nucleated piece of protoplasm of an 
echinoderm ovum with the sperm nucleus of another species ; 1 
development ensued, and the larva resembled the paternal 
form (7). 

In discussing the relations which exist, or are supposed 
to exist, between the cytoplasm and the nucleus, it is clearly 
of the first importance to know what are the changes which 
occur in them, and especially in the nucleus, during the 

1 The animals actually employed were Echinus microtuberculatus 
(male), and Sphaerechinus granulans (female). 


growth, maturity and senescence of the cells. Some ex- 
tremely interesting results in this direction have recently been 
published by Zacharias (8). An ordinary resting nucleus 
consists, as all biologists are aware, of a somewhat dense 
thread-like framework, often spoken of as linin, which 
usually exhibits copious anastomosis, sometimes to such a 
degree that it almost forms a spongy texture. In this 
framework granules are found embedded which react 
definitely to stains and to solvents ; they constitute the 
nuclein, a phosphorus-containing substance which at the 
periods of nuclear division undergoes an enormous increase 
in bulk. The linin is bathed in a more fluid substance, the 
paralinin. One or more spherical bodies, the nucleoli, 
are often present in addition to the foregoing constituents, 
and the nucleus is delimitated from the cytoplasm by a 
pellicle or membrane. The nucleolus contains, as was 
shown by Zacharias many years ago, at least two 
substances, one of which is of an albuminous nature, and is 
dissolved out on treatment with gastric juice ; after peptic 
digestion has extracted the albumin, a substance is left 
which Zacharias calls Plastin. Now observation shows 
that the relative proportion of these two constituents varies 
considerably at different periods of the life of the cell, and 
this is of importance in connection with the intricate series 
of changes which the nucleus passes through during the 
process of ordinary division. The conviction has slowly 
been forced upon us within the last few years that there 
exists a considerable variety amongst the bodies which 
have been included in the common term of nucleoli. 
Auerbach (9) showed in 1890 that some of them 
absorbed certain red dyes with greater avidity than they did 
certain blue ones, whilst other nucleoli reacted in the oppo- 
site manner. He thus distinguished between erythrophil 
and cyanophil nucleoli. These results have been extended 
to plants by the investigations of Rosen (10) and others, 
but especially by Zacharias, who has applied the test of 
solvents to them, with the result that the difference between 
the two classes of nucleoli proves to be a much more real 
one than had hitherto been supposed. And these observa- 


tions are specially interesting when considered from the 
point of view of the great dissentience of opinion which exists 
between most botanists and zoologists as to the nature and 
function of the nucleolus. Strasburger, who admitted the cor- 
rectness of Rosen's statements, considered that the difference 
between an erythrophil and a cyanophil nucleus was largely 
one of nutrition, and he instanced in support of his view 
the difference between the erythrophil nucleolus in the 
nucleus of the well-nourished oosphere and the cyanophil 
nucleus of the much smaller, and therefore presumably 
worse nourished generative cell of the pollen tube. But 
Zacharias, in criticising Strasburger's views, considers that 
there is no evidence to prove that the one nucleolus is in a 
better position than another as regards its nutrition, and it 
is still more difficult to accept the suggested explanation in 
those cases in which both forms of nucleoli are concomi- 
tantly present. 

Zacharias has shown that whereas the erythrophil 
nucleoli contain albumin and plastin, the cyanophil kind 
(the "pseudo-nucleoli" of Rosen and others) contain nuclein, 
a substance quite absent from the other class of nucleoli. 
Rosen in 1892 stated his conviction that his pseudo-nucleoli 
in reality consisted of chromatic substance (nuclein) and 
that they contribute to the formation of those remarkable 
bodies, the chromosomes, which are evolved by the break- 
ing up of the linin framework after the amount of nuclein 
has greatly increased in it, previous to the division of the 
nucleus. Now the nucleolus exhibits striking chancres both 
during the growth, and also during the division of the cell and 
its nucleus. As regards the behaviour during cell growth, the 
relation of the nucleolus to theothercomponents of the nucleus 
is highly suggestive, and seems to support the view of those 
who hold that its function is largely, at any rate, nutritive. 

In the embryonic tissue situated at the growing points 
of plants, the cells are all much alike, differentiation and 
specialisation only taking place behind these regions. 
Consequently it is possible to trace the changes which a 
cell exhibits during its transition from a primitive state to 
its adult form, and often, further, through the various stages 


of senescence and death. Some cells, indeed, are not really 
useful to the plant of which they form a part, until they are 
dead, i.e., till the wall of the cell alone remains, whilst from 
its cavity the protoplasm has disappeared. 

The researches of Zacharias and of Rosen, which have 
recently been published, were directed especially to the 
behaviour of nuclei in the apical regions of plants, and 
their results in the main are confirmatory of each other, 
though the two observers were interested in rather different 
aspects of the same problem. The nuclei of all actively 
dividing cells are markedly cyanophil, and this character is 
especially noticeable just below the active generative cells. 
At first sight it may seem remarkable that in a fern root 
the nucleus of the large apical cell is less cyanophil than 
are the nuclei of the dividing segment cells which have 
been cut off from it. But the anomaly is only apparent, 
for though all the cells in the root owe their origin ulti- 
mately to the division of the apical cell, it must not be 
forgotten that the nuclear divisions in the segments which 
are cut off from it are far more frequent. The segments 
divide up into a very large number of cells before they 
finally form permanent tissue cells, and therefore it is not 
surprising to find that the nucleus of the apical cell, which 
is the ancestor of them all, contains less nuclein than the 
more actively dividing descendants. But there are several 
other significant observations which go to show that in cells 
which are in a state capable of further division, this faculty is 
correlated with the presence of nuclein in their nuclei. Rosen 
found in the roots of the bean and other flowering plants 
that after the tissues were beginning to show differentiation, 
the zone of cells forming the pericycle 1 retained, in their 
nuclei, the characters of embryonic cells, that is to say, 
that, whereas the nuclei of the rest were losing their cyano- 
phil character and were becoming erythrophil, the pericyclic 
•nuclei retained their nuclein contents. . Now the lateral 
roots arise in this pericyclic layer, and they do so by the 
differentiation in it of new growing points. Hence these 

1 A zone of parenchymatous cells sheathing the more central wood and 
bast parts of the vascular strand. 



new rootlets can only be developed from cells which still 
retain, or can re-awaken, embryonic characteristics. Be- 
hind the region in which lateral roots arise, the cells of 
the pericycle lose their cyanophil nature, and here again 
the loss is first apparent in those cells from which, even 
normally, no roots would originate, viz., those situated 
opposite the phloem. It would be interesting to know 
whether in the case of those roots in which the lateral 
rootlets arise right and left of the protoxylem (e.g., Cruci- 
ferse) a corresponding difference obtains. 

Again, Zacharias noticed that during the development 
of the guard-cells of the stomata in a number of leaves 
a similar difference held good. In a simple case, e.g., 
many Liliacese, the mother-cell of the guard-cells is cut 
off from a cell which is destined at once to form one of the 
ordinary and relatively large epidermal cells. In this case, 
whilst the nucleus of the mother-cell of the stoma retains 
its nuclein contents, the other one rapidly becomes poorer 
in this constituent, it grows and develops a large nucleolus. 
The small mother-cell again divides to form the guard-cells 
of the stoma, and only then does a nucleolus become at all 
conspicuous, and the nuclein diminish in quantity. And 
therewith the further capacity for division ceases. 

Besides the connection which is shown to exist between 
a nucleus which is capable of division, and its richness in 
nuclein, there are certain other facts of importance which 
demand notice. The nuclei of cells which are actively 
dividing are commonly characterised by the possession of 
smaller nucleoli than are those in which no further divisions 
will take place, but which are still growing in size. In fact 
Zacharias states generally that, as regards nuclei of cells 
emerging from the meristem region, the nucleoli first 
increase to a maximum, that this is accompanied by an 
enlargement of the nucleus as a whole, which however only 
reaches its maximum size after the nucleolus has done so, 
and that the latter body then diminishes faster than does 
the nucleus as a whole. 

Further, Zacharias found that not only is the nucleolus 
losing substance in those cells which are specialising to 


form tracheids, vessels and sieve tubes, but that the nucleus 
as a whole is losing, and still more rapidly, those substances 
which are capable of being removed by peptic digestion 
from the cell. The facts seem to suggest that it is albumin, 
or some other proteid, which is disappearing ; and it is 
clear that the loss is due to a change in the nucleus itself, 
irrespective of the amount of nutrition available in the 
surrounding plasma, since the change is extremely obvious 
in the degenerating nuclei of sieve tubes, in spite of 
the fact that they are surrounded by abundant albuminous 
substances in the slimy contents of the cells. On the other 
hand, in those cells which are growing in size, preparatory 
to further divisions, such as in spore-mother-cells, the 
increase in albuminous substances, both in the nucleus 
generally, and especially in the nucleolus, is strongly marked. 
Spore-mother-cells, as a rule, pass through a relatively long 
period of growth, and hence we might perhaps anticipate 
(as we find to be the case) that they exaggerate the changes 
seen in the dividing and growing cells of the apical meri- 
stem. But I do not wish to lay too much stress on this, 
because we know that other, and profound, changes occur 
during the growth of spore-mother-cells, and it is uncertain 
to what extent the facts just mentioned may be connected 
with them. 

It may possibly be objected that observations like those 
of Zacharias are open to adverse criticism on the ground that 
the chemistry, and a fortiori the microchemistry, of the 
proteids and other substances which occur in cells is as yet 
in such an unsatisfactory condition. But this objection is 
really not a legitimate one. We know that certain struc- 
tures in the cell are differentiated by their selective action 
on certain dyes, and it is to this fact that their recognition 
was due in the first instance. But we find the action of 
certain solvents to yield no less definite results. Given a 
nucleus in a particular condition (as judged by the structure 
rendered visible by staining), and it will be found that the 
degree of solubility of its constituent substances is charac- 
teristic for the particular stage in the life history of the cell 
or of the nucleus which may happen to have been selected. 


Hence it seems clear that the two methods ought both to 
be employed ; for whilst the staining exhibits more or less 
completely the structural arrangement of the substances 
present, the microchemical method not only indicates some 
at least of the important differences which exist between 
the different structures revealed by the action of staining, 
but it teaches us that certain of these same structures are by 
no means so homogeneous in their nature as one might be 
led to suppose relying on the evidence derived from stain- 
ing alone. 

But those who pin their faith on stains sometimes seem 
to forget that they are after all only employing a sort of 
microchemical method themselves. For the fact that 
different histological elements of the cell are distinguishable 
by stais, implies the existence of a chemical dissimilarity 
between them. And this becomes the more obvious when, 
owing to periodically recurring changes in the cell, we 
assert that this or that structure is growing or diminishing. 
The investigator who is consciously proceeding on micro- 
chemical lines is at least not so open to the charge of mere 
empiricism as are those who look for salvation to 
haematoxylin or the anilin dyes. He may be wrong in 
supposing, for example, that the phosphorus within the 
nucleus only occurs in the nuclein, just as he may be in 
error in assuming that the substance nuclein itself really re- 
presents a chemical substance in the same way that sugar 
does. But he materially advances our knowledge of the 
cell when he determines the fact that a body which fluctuates 
in size as does the nucleolus, is composed of two substances 
or groups of substances one of which is soluble in gastric 
juice whilst the other is not ; and that further, the relative 
size is, in the first instance, correlated with the amount of 
substance which the fermentative action of pepsin can render 

It is readily conceded that the bodies we call nuclein, 
plastin, and the like, possibly may not, as stated already, 
represent chemical molecules at all. This does not, how- 
ever, diminish the interest attaching to the proof that this 
or that substance is at one time present, while at another 


time it can be no longer recognised in its former place. 
Nor does this observation lose in importance when the 
differences are shown to closely accompany changes in the 
general characters of the cells themselves. 


(i) WAGER, H. Preliminary Note on the Structure of Bacterial 
Cells. Annals of Botany, vol. ix. 

(2) Von Sachs. Physiol. Notizen II. Flora, 1892. Also 

Physiol. Notizen IX. Flora, Erganzungs bd., 1895. 

(3) STRASBURGER. Ueber d. Wirkungssphare d. Kerne u. d. 

Zellgrosse. Histologische Beitrdge, v., 1893. 
(3«) STRASBURGER. Schwarmsporenjgameten.Pflanzlichen sperm- 
atozoiden, und das Wesen d. Befruchtung. Hist. Beitr., iv., 

(4) Boulenger, G. A. On the Scaling on the Reproduced Tail 

in Lizards. Proc. Zool. Soc, 1888. 

(5) HABERLANDT, G. Ueb. d. Beziehungen Zwischen Function 

u. Lage d. Zellkerns b. d. Pflanzen. Jena, 1887. 

(6) KORSCHELT. Beitrager 2. Morph. u. Physiol, d. Zellkerns. 

Zool. Jahrb., 1889. 

(7) BOVERI. Ein Geschlechtlich erzeugter Organismus ohne Mtit- 

terliche Eigenschaften. Sitzungsber. d. Gesellsch. f. Morph. 
u. Physiol, zu Miinchen, 1889. 

(8) ZACHARIAS, E. Ueb. d. Verhalten d. Zellkerns in Wachsenden 

Zellen. Flora, Erganzungs bd., 1895. 

(9) AUERBACH, L. Zu Kentniss d. Thier. Zellen. Sitzungsber. d. 

Kgl. Preuss. Akad. d. Wissensch., 26th June, 1890. 
(10) ROSEN, F. Ueb. tinctionelle unterschied verschied. Kernbes- 

tandtheile u. d. sexualkerne. Colitis Beitr. z. Biol. d. Pflanzen, 

v., 1892. 
(iOtf) ROSEN, F. Beitr. z. Kentniss d. Pflanzenzellen. Colitis 

Beitr., vii., 1895. 

J. Bretland Farmer. 


^ > H E recent publication of a number of new manuals and 
monographs dealing with the Mollusca offers a favour- 
able opportunity for a review of our knowledge of this 
great phylum of the animal kingdom. It is not fifteen 
years since Professor Lankester's classical article on Mollusca 
was published in the Encyclopedia Britannica, yet the con- 
tributions to Molluscan morphology since that date have 
been not only numerous, but in many cases of prime im- 

The older method of inquiry, that of the comparison of 
types more or less arbitrarily selected from different groups, 
has been succeeded by investigations more directly in- 
fluenced by the idea of evolution. The comparison of types 
has been replaced by the study of groups. The founda- 
tions of the morphological edifice were laid upon the former 
method ; the superstructure and details are the result of 
the latter. Homologies having been to a large extent 
determined, we now seek phylogenies. It happens also 
from time to time that the detailed study of a group with 
the object of reconstructing the phylogeny of its members 
leads occasionally to the discovery that homologies based 
on the simple method of anatomical comparison turn out 
to be nothing more than analogies — recurrent examples of 
similar modifications. 

One result of these phylogenetic inquiries has been the 
concentration of particular attention upon forms which are 
presumably the most primitive in each group ; and great 
advances have thus been made in our knowledge. Kow- 
alewsky and Marion, Pruvot, Wiren, and Thiele have 
enormously extended our acquaintance with the Apla- 
cophorous Isopleura ; primitive Prosobranchs (Docoglossa 
and Rhipidoglossa) have been thoroughly investigated by 
Haller and Boutan ; Bouvier has thrown new light upon the 
Opisthobranchia by his researches on Actceon ; Boas and 
Pelseneer have revolutionised our ideas of the Pteropoda 


by their work upon Limacina among the Thecosomata, and 
upon Dexiobranchcea and other types among the Gymnoso- 
mata ; the morphology of the Pelecypoda has been further 
elucidated by Pelseneer's observations upon Nucula and 
other primitive forms, and important contributions to our 
knowledge of the Cephalopoda were made during the past 
year by Huxley and Pelseneer in the case of Spirilla, that 
last survivor of the ancient types of Decapod Dibranchiates. 
We doubt if any equivalent group of the animal kingdom, 
except perhaps the Echinoderma, has been the subject of 
such productive researches as the M ollusca during the period 
under consideration ; and certainly the phylogenetic method 
of inquiry has attained no greater triumphs than in the 
hands of Bouvier, Haller, Pelseneer, and other inves- 
tigators of the Gastropod and Lamellibranch series. 

In the present article I propose to deal more especially 
with recent contributions to our knowledge of the Molluscan 
nervous system, reserving a fuller consideration of other 
questions for a later article. 

There is one writer, however, whose views must first of 
all be dealt with, as on a great number of fundamental 
points they are opposed to all current conceptions of 
Molluscan morphology. These views merit some detailed 
consideration, moreover, for they are based on propositions 
which are not without a certain appearance of plausibility, 
and may well serve as test-questions by which to examine 
into the accuracv of the homologies which have been 
generally admitted to exist between the different sections 
of the Molluscan phylum. 

Thiele has published his views in a series of lengthy 
papers, the references to which will be found in the biblio- 
graphy (23, 24, 25). He regards the Mollusca and Anne- 
lida as direct descendants of Polyclad Turbellarians, and 
his identifications of homologous organs in the different 
Molluscan groups are determined, not by a direct comparison 
of the organisation of these types one with another, but 
by independent comparisons of the organisation of the 
different Molluscan types with that of sucker-bearing 
Polyclads. The group Mollusca is thus made to lose its 


compactness, and characteristic organs, such as mantle and 
ctenidium, which have been regarded as homologous 
throughout the Molluscan series, are interpreted in different 
ways in the different types, as the exigencies of Thiele's 
theory demand. One of the first propositions assumed by 
this writer is that the foot of the Mollusca is simply a colossal 
enlargementof the ventral sucker of the Polyclad; thesuctorial 
function of the foot in Chiton and the lower Gastropoda is 
pointed to in support of this comparison. A series of 
more revolutionary propositions is then promulgated in 
consequence of the necessity under which the author is 
placed of discovering the primitive body-edge of the 
Mollusca comparable to the edge of the body of the Tur- 
bellaria. This primitive body-edge Thiele identifies by 
means of the lateral sense-organs which characterise the 
epipodium in the Rhipidoglossa and the margin of the 
mantle in Pelecypoda. The epipodium in Gastropoda and 
the mantle edge in Pelecypoda are thus taken by this writer 
to represent the sides or edge of the body in the Tur- 
bellarian ancestor. The epipodium in Gastropoda and the 
mantle edge in Pelecypoda consequently separate the 
dorsal from the ventral regions of the body in those groups. 
It follows from this that the ctenidia of Gastropoda, which 
are supra-epipodial in position, are not homologous with 
the ctenidia of Pelecypoda, which are infra-pallial. How we 
are to regard the anus, which is dorsal in the one group and 
ventral in the other, is not explained. But since in oper- 
culate Rhipidoglossa the operculum, like the shell, is 
situated above the epipodium, we are told that the oper- 
culum must also be regarded as dorsal in position, as well 
as serially homologous with the shell proper. This, in 
Thiele's eyes, compares well with the condition of affairs in 
Chiton, whose shelly plates are without doubt serially 
homologous. Moreover, although the existence of an 
epipodium in Chiton has not been hitherto recognised, 
Thiele argues that, since the pallial fold in this form re- 
presents the primitive body-edge, it must also, together with 
the series of ctenidia which are attached to its lower surface, 
be regarded as the homologue of the epipodium of the 


Rhipidoglossa. The ctenidia of Chiton are, in fact, re- 
garded as modified epipodial cirri. The consequence of 
this view is that while the mantle of Chiton and the mantle 
of Pelecypoda are regarded as homologous, the mantle of 
the Gastropoda is supposed to represent only a portion of 
the mantle in these other forms, and its projecting rim, 
similar as it appears to be in the two cases, is held to be a 
new and secondary formation unrepresented in the Am- 
phineura and Pelecypoda. 

Nowhere, however, do we find in Thiele's voluminous 
writings any explanation of the anomaly which ought to 
have occurred to him, that while in Chiton the anus is 
"ventral," and lies well beneath the "epipodium" and the 
last shell-plate, in operculate Gastropods the intestine opens 
not only above the epipodium, but between the operculum 
and the shell of the embryo — a relation which could only be 
represented in Chiton, if Thiele's theories were correct, by 
the situation of the anus between two of the shell-plates 
upon the back of that animal ! 

The nervous system of the Mollusca is treated by Thiele 
with a ruthlessness no less than that which is meted out 
to the external organs of the body. Let us take the 
Amphineura first. In this group, if the relations of the 
nervous system in Chiton be taken as typical, we have 
dorsal to the gut a great ganglionic nerve-ring whose lateral 
components are usually referred to as the lateral or pleuro- 
visceral cords. Connected anteriorly with the cerebral 
enlargements of this nerve-ring is a pair of ventral or pedal 
cords, connected with one another by a series of commis- 
sures lying beneath the gut, and also with the lateral cords 
by means of lateral connectives. The lateral cords inner- 
vate the pallial sense-organs, gills, and viscera ; the ventral 
cords the musculature of the foot. The lateral cords are 
regarded by Thiele as the homologues of the lateral cords 
or nerve-ring of the Turbellarians. and the ventral cords are 
taken to correspond to the ventral longitudinal nerves of 
the same forms. So far we find nothing either erratic or 
original, for the same view has already been taken by Lang 



But the novelties begin with Thiele's interpretations of 
the nervous system of Gastropoda and Pelecypoda. We 
have already pointed out Thiele's view that the epipodium 
of Gastropods represents the primitive body-edge. Now 
at the base of the epipodium in Fissurella and Haliotis there 
lies a ganglionic plexus ; and this plexus, which takes the 
form of an incomplete ring, is regarded as the homologue 
of the lateral cords of Turbellarians and Amphineura. The 
series of epipodial nerves which connect the epipodial plexus 
with the upper half of the pedal cords in Rhipidoglossa is 
compared with the series of connectives between the lateral 
and ventral cords in Amphineura. 

This seems very plausible until one recollects (i) that, 
the epipodium being infra-rectal, the epipodial plexus is 
also infra-rectal and thus difficult to compare with the 
lateral cords of Amphineura, whose "commissure" is supra- 
rectal ; and (2) that, whereas in Amphineura the lateral 
cords innervate practically the whole of the pallium and 
viscera, in Rhipidoglossa the epipodial plexus has nothing 
to do with any other organs except the sense-organs of the 
epipodium. If the pallium of the Gastropoda is really, as 
Thiele maintains, a secondary differentiation of the primary 
pallium of the Amphineura, one would expect that its 
innervation would also be effected by progressive differen- 
tiation of the nerve-centres which supplied the primary 
pallium, viz., from the lateral or epipodial centres. So far 
from this being the case, however, Thiele himself (xxv., pp. 
587-9) adopts the view that the pallial nerves as well as the 
pleural ganglia of Gastropoda are secondary derivatives of 
the ventral or pedal cords. 

The recklessness of Thiele's comparisons reaches its 
high-water mark, perhaps, in his remarks on the nervous 
system of Pelecypoda. Correlated with the existence of 
numerous sense-organs (eyes, tentacles, etc.) along the 
mantle edge, there exists in many forms {Area, Pecten, 
Pinna, etc.) a nervous ring around the mantle which may 
take the form either of a complete ring of peripheral ganglia 
united by a plexus, or of a circumpallial ganglionated nerve, 
as was recognised by Duvernoy (5) more than thirty years 


ago. Since the mantle-lappets of the two sides of the body 
unite posteriorly above the anus, this pallial nerve-ring lies 
above the gut. The ring is connected with the cerebro- 
pleural ganglia by means of the anterior pallial nerves, and 
with the visceral (parieto-splanchnic) by means of branches 
from the great posterior pallial nerves. Accordingly Thiele 
homologises the circumpallial nerve-ring with the lateral 
cords of Chiton and with the epipodial plexus of the Rhi- 

The first of these homologies seems not unreasonable, for 
no one disputes the homology between the mantle of Chiton 
and that of Pelecypoda. Moreover Kowalevsky's discovery 
that Chiton in its later embryonic phases is provided with 
a pair of transitory eyes which lie outside the velar area 
and have some close connection with the lateral nerve- 
cords, renders this comparison particularly worthy of 
attention. But how the circumpallial nerve of Pelecypoda 
can be in any sense homologous with the epipodial plexus 
of Gastropoda, when the latter structure lies beneath the 
gut and has no connection with the cerebral ganglia, either 
directly or by the intermediation of the pleural ganglia, it 
is altogether impossible to conceive. And this is not all. 
The posterior connection between the circumpallial nerve 
of Pelecypoda and the visceral ganglia is compared by 
Thiele with the posterior connectives between the lateral 
and ventral cords of Amphineura ; and the time-honoured 
visceral nerve-cords of Pelecypoda, with the visceral (parieto- 
splanchnic) ganglia upon them, are homologised with the 
ventral cords of the Amphineura. To reveal the absurdity 
of these comparisons it is sufficient, I think, to remind my 
readers that the ventral cords of Chiton are concerned ex- 
clusively with the innervation of the musculature of the 
foot ; while the visceral cords of Pelecypoda innervate 
the body-wall, ctenidia and viscera^ in addition to the 
posterior adductor muscle. How these supposed homo- 
logues of the ventral cords of Chiton have come to assume 
so many of the functions of the lateral or pallio-visceral 
cords, is not explained ; and since Pelecypoda possess a 
pair of pedal ganglia in the foot, as typical in their relations 


as those of any Gastropod — in Nucula to the extent even 
of having separate cerebro-pedal and pleuro-pedal connec- 
tives (18, 19) — it seems profitless to pursue these ill- 
balanced speculations any further. 

The utmost ingenuity cannot overcome the fact that 
there is a fundamental disparity between the Turbellarian 
and Molluscan body. This disparity is revealed by em- 
bryology ; but to embryology Thiele pays scant attention. 
Thiele's argument is practically this (24, p. 504), — that 
the only route from Ccelenterates to Bilateralia is via the 
Ctenophores to Polyclads, and that Annelids and Molluscs 
are consequently to be derived from Polyclad ancestors. 
Embryology seems to me, however, to point to two lines of 
descent at least, from the Ccelenterates to the Bilateralia. 
In each case the oral surface of the Ccelenterate ancestor 
became the ventral surface of the Bilateral descendant ; but 
along one line of descent the primitive mouth or blastopore 
retained its ancestral form as a simple circular orifice in the 
middle of the ventral surface, and opened into a gastral 
cavity devoid of an anal orifice (Polyclads) ; while along the 
line of descent which led to the Annelida and Mollusca the 
blastopore elongated along the ventral surface, as Sedg- 
wick has so ably contended, its lips coalesced except at the 
two extremities, and these open ends constituted the mouth 
and anus of the Ccelomate descendants. Thiele has 
altogether overlooked the significant behaviour of the blas- 
topore in Annelidan and Molluscan embryos ; and since 
no similar modification of the blastopore is known in the 
case of Turbellarians and Trematodes, in which groups the 
absence of an anus is so marked a characteristic, we are 
amply warranted, I think, in drawing the conclusions which 
I have emphasised above. 

The admission of this distinction is however fatal to 
any theory of the Polyclad ancestry of the Mollusca. The 
foot of the Mollusca is a development of the fused lips of 
the elongated blastopore, and can in no case be homo- 
logised with the ventral sucker of Turbellarians which lies 
entirely behind the blastopore. The same remark applies 
to Lang's comparison of the Molluscan foot with the ventral 


surface of the Turbellarian. The foot is undoubtedly part 
of the ventral surface of the Mollusc, and as such may be 
compared, in a general way, with the creeping surface of a 
Planarian ; but as a specialised organ, developed from the 
fused lateral margins of a slit-like blastopore, it has no 
homoloeue in the organisation of the Turbellaria. 

Let us now see what light has been thrown on the 
problems of Molluscan morphology by the researches of 
other investigators. 

The visceral commissure. — One of the greatest dif- 
ficulties in comparing the Amphineura with the Gastropoda 
or other Molluscan types has long been the fact that the 
lateral or pleuro-visceral cords of Chiton, which innervate 
the gills, viscera, and mantle, are united to one another 
posteriorly by a "commissure" lying above the rectum; 
whereas the visceral commissure of Gastropoda and Pelecy- 
poda, etc., lies below the intestine. 

A little care in the use of words would have prevented 
much of the confusion and controversy which has arisen on 
this subject of the position of the visceral commissure. 
Words, as Bacon phrases it, put constraint upon the in- 
tellect, and there is no doubt that the disagreement and 
perplexity of naturalists concerning this point have been 
caused by one of the idola fori which they have themselves 
set up, rather than by any intrinsic incompatibility in the 
facts themselves. If the language must still be maintained, 
I must at least point out that there are commissures and 
commissures, and that one may be a commissure in fact, 
and another only in name. The suprarectal ''commissure" 
in Amphineura is ganglionic, and, like the rest of the 
pleuro-visceral nerve-ring, is formed in situ by delamination 
from the ectoderm (15). It is not a commissure in the 
strict sense of the word, but an integral portion of an 
annular central nervous system. But the visceral loop of 
other Molluscs consists merely of nerve-fibres connecting 
usually a couple of visceral ganglia with one another, and 
with the pleural ganglia. Now nerve-fibres are outgrowths 
from nerve-cells, and if two groups of nerve-cells should 
happen to take a somewhat deep-seated position in the body 


before their fibres have grown out (which is not a rare 
embryological phenomenon), there should be nothing in- 
comprehensible in their fibres taking the shortest route and 
meeting beneath the gut instead of over it. Clearly, there- 
fore, the ventral position of the visceral commissure in 
most Mollusca by no means precludes the possibility of the 
essential homology between the visceral loop of these 
forms and part of the pleuro-visceral ring of Amphineura. 

The other differences between the visceral loop of most 
Mollusca and the pleuro-visceral ring of Amphineura are- 
principally differences in the degree of segregation and 
concentration of ganglion-cells and nerve-fibres. The 
pleuro-visceral ring of Chiton represents a very primitive 
nervous system, characterised by the more or less even 
diffusion of ganglion-cells over the whole length of the 
cord, while the nerves arising from it are not united into 
large trunks, but are given off at repeated intervals in a 
manner which is almost metameric. The nerves springing 
from it innervate the same parts of the body as the com- 
bined pleural and visceral ganglia of Gastropods and other 
Molluscs, viz., mantle, ctenidia, intestine, heart, nephridia, 
and gonads. But if, after the reduction of the ctenidia to 
a single pair, we imagine a process of segregation to set in 
between these various elements, the more strictly visceral 
centres would become separated from the superficial pallial 
centres, and would assume a deeper position in the body. 
The law of concentration would apply in this as in other 
cases of evolution of nervous systems (3), and the result of 
the whole process would be the differentiation of a visceral 
nervous system, consisting of ganglia and commissural 
fibres, out of the primitively mixed and diffuse pleuro-visceral 
system. If the primitive relations to the gut and ring-like 
form were retained at all, they would be retained, not 
necessarily by the visceral system, which has ex hypothesi un- 
dergone considerable changes, but by the pallial (= pleural) 
system, which has undergone no change, except possibly 
one of incipient concentration. 

The position of the commissural fibres of the visceral 
ganglion in relation to the gut becomes a matter of sub- 


ordinate importance if the evolution of the nervous system 
has proceeded upon these lines, as will be made evident 
later on. As a matter of fact the visceral commissure is 
situated below the gut — a relation which is possibly fore- 
shadowed in Chiton by a connection beneath the gut of the 
two gastric nerves described by Haller (8). 

Pelseneer (19) indeed goes so far as to identify these 
gastric nerves of Chiton with the visceral commissure of 
Gastropoda and Pelecypoda; but the considerations which I 
have emphasised above show that the typical visceral nerves 
and commissure have not yet arisen in the Amphineura ; 
they do not arise, in fact, until the branchial, nephridial, 
genital and enteric branches of the primitive pallio-visceral 
cords are all united into one common trunk. There is 
some doubt, moreover, as to the existence of the gastric 
nerves described by Haller, since two investigators, Plate 
(20) and Thiele, have been unable to discover them in 
species of Chiton examined by themselves. 

A valuable contribution to this part of the subject is 
contained in Haller's recent Studien (11). In the 
common cyclobranchiate types of Limpet the pallial nerves 
are separate from one another behind, and seem to be 
mere outgrowths of the pleural ganglia (Bouvier, 3, p. 19); 
but in Lottia, one of the more primitive monobranchiate 
forms, Haller shows that the pallial nerves of the two sides 
are directly continuous with one another posteriorly, and make 
a complete arch round the edge of the mantle. They are 
moreover not mere nerves, since they consist of a core of 
fibres surrounded by an outer coating — discontinuous, it is 
true — of ganglion-cells. They are clearly the posterior 
continuations of the pleural ganglia, and represent the re- 
mainder of the pallio-visceral nerve-ring of the Amphineura 
after the separation of the visceral elements. This view is 
further borne out by the existence of several connectives 
between the pallial ring and the pedal cords in addition to 
the stout ganglionic connective which in higher forms 
becomes the persistent pleuro-pedal connective. 

The pleural ganglion. — Haller's discovery recorded in 
the preceding paragraph shows clearly the error of the 


view by which the pleural ganglion is regarded as a 
derivative of the pedal cords (Bouvier, Pelseneer, etc., 
passim). This view is founded on the fact that in the lower 
Gastropoda (Docoglossa and Rhipidoglossa) the pleural 
ganglia are directly continuous with the anterior ends of 
the pedal cords, while in the higher types the pleural ganglia 
gradually move further and further away from the pedal 
ganglia, and, travelling along the cerebro-pleural connectives, 
eventually come into contiguity with the cerebral ganglia 
(Tenioglossa) or even fuse with them to form a single 
cerebro-pleural ganglion on each side (Pelecypoda). 

The close connection between the pleural and pedal 
ganglia in the lower forms may now be interpreted 
in a different manner. The ganglion-cells which were 
primitively distributed over the whole extent of the pallial 
nerve-ring have been concentrated at the anterior ex- 
tremities of its lateral portions, as Haller's observations 
on Lottia show— or rather in the reo-ion of the first 
pleuro-pedal connective, for the most anterior portion of 
the primitive pallial cords is represented by the cerebro- 
pleural connective. The shortness of the pleuro- 
pedal connecting piece and the great concentration of 
ganglion-cells which takes place at its two extremities 
prevent any sharp demarcation between the pleural and 
pedal ganglia in these lower forms ; but a comparison of 
the nervous system of Lottia with that of Chiton (Thiele, 
2 3 ? P- 387) leaves no room for doubt as to the correct- 
ness of this interpretation, which throws a flood of light 
upon numerous other points which have been difficult to 
understand upon the older views. It explains, for example, 
why the cerebro-pleural and cerebro-pedal connectives 
should be already distinct from each other in the lower 
Gastropods at a stage when the pleural ganglia are in 
actual continuity with the pedal cords, and it sets at rest 
the controversy as to the meaning of the lateral furrow 
in the pedal cords of Rhipidoglossa which has been waged 
with so much skill in the rival pages of the Archives de 
Zoologie and the Bulletin Scientifique de la France et de la 


Development of the pleural ganglion. — That the pleural 
ganglion is essentially distinct from the pedal is, I think, 
sufficiently clear from the facts of development. Although 
these ganglia are placed so close together and are so inti- 
mately connected in the lower Gastropods there is not a 
single case on record in which the pleural ganglion has been 
observed to arise from the pedal ganglion, or from a 
common pleuro-pedal rudiment in the embryo. It is 
equally true on the other hand that Sarasin's derivation of 
the cerebral and pleural ganglia from a common rudiment 
in Bithynia (the cephalic sense-plate) has been opposed by 
v. Erlanger, who shows that all the great ganglionic centres 
arise separately, and do not become connected with one 
another until after their differentiation (7). 

A renewed investigation of the origin of the cerebro- 
pleural ganglion in Pelecypoda would be of great interest 
in this connection. Pelseneer's ( 1 8) observations on Nuctila 
have placed the fact of the composite nature of this ganglion 
in Pelecypoda beyond all doubt ; and still, to the best of 
my knowledge, no one has yet observed the appearance in 
the embryo of a pleural element distinct from the main body 
of the ganglion. This apparent community of origin of the 
cerebral and pleural ganglia in Pelecypoda may be compared 
with the direct continuity of the cerebral and pleural 
elements of the nervous system in Amphineura. 

Development of the visceral ganglia. — Sarasin en- 
deavoured to show that the visceral ganglia of Bithynia, 
together with the pedal and abdominal ganglia, arise in the 
embryo from a common ventral proliferation of the ectoderm 
which he compares with the ventral ganglionic chain of 
Annelida. On this point also Sarasin has been corrected 
by v. Erlanger, who shows that all these ganglia arise 
separately from one another in Bithynia (7), as well as in 
Palndina (6). 

The visceral ganglia are also quite distinct from the 
pleural ganglia in their origin, as v. E Hanger's observations 
show. In one important respect, however, the visceral 
ganglia and the pleural ganglia betray a marked similarity, 
the significance of which seems, however, to have escaped 



the attention of its discoverer. In Paludina v. E danger 
figures the pleural ganglia arising from the ectoderm on 
each side of the body at a point just outside the velar area, 
but in actual contiguity with the cells of the ciliated ring. 
In Bithynia (7, Taf. xxvi., fig. 16) he figures the same 
condition of things for the pair of visceral ganglia. The 
only difference in origin between the two ganglia is that the 
visceral ganglia arise behind the pleural ganglia. If the 
Molluscan veliger possessed a nerve-ring beneath its proto- 
troch (velum), as occurs in the trochosphere of the Annelida, 
it is quite clear that the pleural and visceral ganglia of 
Bithynia and Paludina would represent a series of gangli- 
onic thickenings along the course of the nerve-ring. Apart 
from this inference, however ; the topographical relations to 
which I have called attention seem sufficient to establish 
the proposition that the pleural and visceral ganglia, and, 
as I shall show directly, the abdominal ganglion also, of 
Gastropods — and, therefore, of other Mollusca — belong to 
a group of dorso-lateral nerve-centres quite distinct from 
that which is represented by the ventral or pedal cords. 
Here again we are reminded of the direct continuity of the 
pleural and visceral nerve-centres in the Amphineura. 

Development of the abdominal ganglion. — In Chiton, 
as Kowalevsky has shown (15), the unpaired abdominal 
ganglion, or, as it is often called, the visceral ganglion, 
arises by a proliferation of the ectoderm at the hinder pole 
of the embryo, dorsally to the site of the future proctodeum. 
In the adult this ganglion is simply a special concentration 
of ganglion-cells on the supra-anal portion of the pleuro- 
visceral ring. 

The abdominal ganglion of Gastropods is also situated 
at the hinder end of the visceral loop, but lies of course 
ventral to the gut. Can these two ganglia be regarded as 
homologous ? 

If Molluscs were mere mechanical models the answer 
would be undoubtedly in the negative ; but embryology 
points unhesitatingly to the opposite conclusion. Von 
Erlanger has shown that in Bithynia as well as in Paludina 
the abdominal ganglion develops as an ectodermal pro- 


liferation of the floor of the mantle-cavity, i.e., that the 
ganglion is essentially a dorsal ganglion. Its final situation 
on the course of the sub-intestinal nerve-loop is rendered 
possible by the fact that its connectives with the visceral 
eanolia are not delaminated from the ectoderm, as are the 
ganglionic pleuro-visceral cords of Chiton, but are mere 
fibrous outgrowths from the ganglia themselves. Embry- 
ology is thus in complete accord with the views which have 
been maintained in the earlier part of this paper as to the 
homologies and origin of the visceral nervous system in 

The pallial and visceral commissures in Cephalopoda. 
— It has long been known (Hancock) that in many Cepha- 
lopoda the stellate ganglia on the pallial nerve-cords are 
connected with one another above the gut by a transverse 
commissure. Is this commissure a relic of the pallio- visceral 
nerve-ring of the Amphineura and homologous with the 
pallial ring of Lottia, or is it merely a secondary connection ? 

In Spirilla a remarkable arrangement of the pallial 
commissure has been recognised by Huxley and Pelseneer 
in their recent memoir (12). The commissure is not in 
this case a straight transverse band, but consists of two 
curved cords which arise from the right and left stellate 
ganglia respectively, and at their junction in the median 
line of the body give off a median pallial nerve which runs 
for a short distance forwards, and then passing over the 
anterior margin of the shell — which is, of course, internal — 
becomes recurrent and runs along the part of the mantle 
contained within the last chamber of the shell. Pelseneer is 
thus led to regard the commissure with its median nerve as 
formed by the two original pallial nerves fused together. 
The connection between the stellate ganglia having thus 
arisen in the primitive Dibranchiates (apparently in con- 
nection with the reduction in size and enclosure of the 
chambered shell), higher forms show a series of stages in 
its subsequent degradation, until it is finally lost in the 
Octopoda. The absence of a pallial commissure in Nautilus 
also supports Pelseneer's view that in Cephalopoda this 
structure is not of any primary importance. 


At the same time when Pelseneer added a paragraph to 
the effect that the supra-rectal commissure of the Amphi- 
neura is also a merely secondary junction of the pallial 
nerves, he was probably not yet acquainted with Haller's 
work on Lottia, and allowed his views upon the Polychsete 
ancestry of the Mollusca to bias his interpretation of the 
Molluscan nervous system. 

In a recent paper on the anatomy of Nautilus Mr. 
Graham Kerr (13) also refers to the question of the supra- 
rectal commissure. It will be remembered that in Nautilus 
the pleuro-visceral ganglia of the two sides form a stout 
ganglionic band encircling the oesophagus in the region of 
the cerebral ganglia. The pallial nerves radiate from the 
lateral portions of this half-ring, and the pair of visceral 
nerves arise from the ventral portion. The visceral cords 
pass backwards on either side of the vena cava, and, after 
giving off the branchial nerves, are prolonged posteriorly as 
far as the post-anal papilla, behind which Mr. Kerr has 
recognised an apparent anastomosis. Mr. Kerr adds that 
in this case " the homologue of the pleuro-visceral cord of 
Chiton is not merely the posterior sub-cesophageal nerve- 
mass, but rather the two lateral portions of this, together 
with the post-branchial prolongations which run on either 
side of the vena cava. The mesial part of the posterior 
sub-cesophageal nerve-mass would therefore be a secondary 
fusion between the nerve-masses of the two opposite 

In his suggested homology of this possible post-anal 
{i.e., supra- rectal) commissure of the visceral nerves in 
Nautilus with the supra-rectal " commissure " of Chiton, 
Mr. Kerr has undoubtedly failed to appreciate the true 
nature of the posterior sub-cesophageal loop of Nautilus, as 
well as the relation of the visceral nerves to the pleuro- 
visceral cords of Chiton. The explanation of the Cephalo- 
pod nervous system is most readily found by comparing it 
with that of Dentalium, whose organisation in many respects 
supplies connecting links between that of the Cephalopoda 
and that of the primitive prae-torsional Gastropod or 
primitive Pelecypod. In Dentalium (22, p. 401) we find 


a pair of post-anal prolongations of the visceral nerves 
precisely resembling those described by Kerr in Nautilus ; 
yet in Dentalium, owing to the smaller degree of concen- 
tration or cephalisation which has taken place in the 
nervous system, it is easy to see that the typical sub-intes- 
tinal visceral commissure exists as in Gastropods and 
Pelecypods. The posterior sub-cesophageal nerve-mass 
of Cephalopods has clearly been produced, not, as Mr. 
Kerr suggests, by a secondary fusion of the pleuro- visceral 
nerve-masses of the two opposite sides, but by a simple 
shortening of the visceral loop as it occurs in Dentalium. 
This would bring the visceral ganglia into continuity 
with the pleural ganglia and with one another, — a process 
of condensation with which we are already familiar in the 
Tenioglossa and the Euthyneura among Gastropoda. 

It may here be mentioned that Willey's simultaneous 
account (26) of the visceral nerves of Nautilus, while con- 
firming Mr. Kerr's observations as to the existence of post- 
anal prolongations of a pair of visceral nerves, differs from 
his statement as to their origin. Willey states that the 
nerves supplying the post-anal papilla arise independently 
from the sub-cesophageal visceral loop, although at their 
origin they are adjacent to the branchial nerves and for a 
large part of their course are actually contiguous with them. 
The significance of this separation is not remarked upon by 
Willey ; but if the separation really exists it is certainly a 
difficulty in the way of his contention that the post- 
anal papilla represents an approximated posterior pair of 
branchial sense-organs, since the anterior osphradium and 
both gill-plumes are all innervated from the outer visceral 

Etithyneurism. — Since the publication of Spengel's paper 
on the olfactory organ and nervous system of Mollusca, a 
division of the Gastropoda into two groups, the Strep- 
toneura and the Euthyneura, has been generally adopted. 
This classification has been accepted, moreover, not merely 
as an expression of the anatomical facts concerning the 
condition of the visceral loop in the two groups, but as a 
classification of phylogenetic significance. It is to be in- 


ferred that the two groups have been independently derived 
from a common type of archi-Gastropod, possessing an un- 
twisted visceral loop — the Prosobranchs (Streptoneura) by 
the twisting of the loop, the Opisthobranchs and Pulmonates 
(Euthyneura) by the mere shortening and concentration of 
the untwisted loop. This view derives support from the 
fact that the persistent ctenidium retains its primitive posi- 
tion on the right side of the body in Opisthobranchs, while 
in Prosobranchs it shows a marked displacement and lies 
on the left side. Bouvier's observations on Actceon 
( = Tornatella), however, have completely altered the posi- 
tion of affairs. Actceon is a very primitive Opisthobranch, 
as may be inferred from the high development of its shell, 
the persistence of its operculum, and the absence of pleuro- 
podial fins. Bouvier tell us (4) that Actcson resembles the 
Prosobranchs, not only in these points, but also in possess- 
ing a distinct twist of the visceral loop (streptoneurism, 
chiastoneurie). The ctenidium is innervated from a supra- 
intestinal ganglion, which lies on the left side of the body. 
We are accordingly led to the conclusion that the 
euthyneurous condition of Opisthobranchs and Pulmonates 
has not been directly inherited from the orthoneurous 
ancestors of the Gastropoda, but has been derived from a 
previously streptoneurous condition. In other words the 
Opisthobranchs and Pulmonates have descended from 
Prosobranch ancestors, and the right-sided position of the 
gill-plume in Opisthobranchs is not primitive, but the result 
of a secondary process of detorsion. 

Orthoneuroidism. — Without going further into the 
matter it may also here be mentioned that the supra-in- 
testinal commissure has been recently discovered in various 
species of Nerita, Neritina, and Navicella by Boutan (2), 
Bouvier (3^), and Haller (11) — a discovery which de- 
stroys the last refuge of orthoneurism in Prosobranchiate 
Gastropods. Streptoneurism may now be affirmed of all 
Prosobranchiate Gastropods. 

Origin of the Moliuscan nervous system. — The attempts 
of previous writers to explain the relations of the nervous 
system of Mollusca have been based almost exclusively 


upon comparisons with the fully constituted nervous 
systems of such types as the Turbellaria and Annelida. 
With Thiele's theory of the Turbellarian ancestry of the 
Mollusca I have already dealt, and I do not propose to 
deal with the Annelidan hypothesis, since this theory can- 
not provide any satisfactory explanation of the high develop- 
ment of the pleuro-visceral nervous system of the Mollusca. 
Those authors who, like Thiele and Pelseneer, homologise 
both the pleural and pedal centres of the Mollusca with the 
ventral cords of Annelids, base their view upon the sup- 
posed origin of the pleural centres from the pedal cords. 
This derivation I have already shown in this article to be 
completely erroneous. Pelseneer's theory of the origin of 
the Mollusca from Polychsete ancestors (18a), and all 
theories which seek the origin of the Mollusca in the 
specialised representatives of any of the vermiform groups, 
may at once in my opinion be dismissed from considera- 

Apart from matters of minor importance it will, I think, 
be conceded that the following cardinal points in regard 
to the morphology of the Molluscan nervous system have 
been established by the facts and arguments which have 
been presented in this article : — 

(1) That the pleural ganglia have not been derived by 

segregation from the ventral or pedal cords. 

(2) That the pleural, visceral, and abdominal ganglia 

of Gastropoda form a group of dorsal nerve- 
centres — the two former owing to their dif- 
ferentiation in the immediate neighbourhood of 
the velum, and the latter owing to its differentia- 
tion from the mid-dorsal wall of the body (floor 
of mantle-cavity). 

(3) That the dorso-lateral nerve-ring of Amphineura 

is primitive and is represented in other groups 
of Mollusca by both the pallial and visceral 
nerve loops, or their derivatives. 

(4) That the sub-intestinal position of the visceral loop 

in all groups except the Amphineura is a 
secondary one, which has been rendered possible 


only by the decentralisation of the primitive 
pleuro-visceral nervous system, and its separa- 
tion into special ganglia and nerves, the latter 
being formed ontogenetically as fibrous out- 
growths from the ganglionic centres. 
Venturing now, in conclusion, upon more speculative 
ground, I believe that the embryonic relations, to which I 
have drawn attention, between the pleural and visceral 
ganglia and the ciliated band are of phylogenetic importance. 
It has long puzzled me that the larval forms (trochospheres) of 
two groups so closely allied as the Annelida and Mollusca, 
while presenting a close similarity in general structure, 
should differ so remarkably in regard to their nervous 
system. The Annelid trochosphere has a nerve-ring 
beneath its ciliated band, while the Molluscan trocho- 
sphere has none. In this respect the Molluscan trocho- 
sphere appears to be less primitive than that of the 
Annelida. The explanation of this now appears to me to 
be as follows. In the evolution of the Annelida the proto- 
troch and nerve-ring remained for a long time unmodified, 
and did not share in the elongation of the postero-ventral 
region of the body which gave rise to the trunk of the 
Annelid. This would explain the absence of the dorsal 
nerve-ring in the adult Annelid, provided that the nerve- 
ring, together with the prototroch, came to have merely a 
larval significance, — as actually happens in the ontogeny of 
Annelids to-day. On the other hand, in the evolution of 
the Mollusca from the same simple type of ancestor, the 
whole body must have shared in the elongation — the proto- 
troch and nerve-ring as well as the more ventrally placed 
parts of the body. This elongated nerve-ring I identify 
with the pleuro-visceral ring of Amphineura, although the 
phyletic connection between the nerve-ring and the ciliated 
band is inferred from the development of certain Gastro- 
pods rather than from the Amphineura themselves. As a 
larval adaptation for conveniences of natation I imagine 
that a separation became gradually effected in embryonic 
life between the ciliated ring and the nerve-ring, the former 
becoming restricted to the anterior end of the larval body, 


while the latter became more and more extended pari passu 
with the elongation of the trunk. Such a separation is to 
some extent paralleled in the development of Holothurians 
from the Auricularia larva, as described by Semon. On 
this theory alone can I explain to myself the absence of the 
ancestral nerve-ring in the trochospheres of Mollusca, and 
I find some support for this view in the ontogeny of Nemer- 
tines. The lateral nerve-cords in this group have the same 
relation to the gut and brain as have the pleuro-visceral 
cords of Chiton, since they form a dorso-lateral ring, the 
posterior commissural portion passing above the rectum. 
In Nemertines there can be very little doubt that this 
nerve-ring has been derived phyletically by the elonga- 
tion of a nerve-ring which underlay the ciliated band of a 
more or less Pzlidzum-like ancestor, as it underlies the 
ciliated band of the Pi/idzum-laxva., although this phyletic 
origin is disguised by the profound metamorphosis which 
breaks the continuity of the ontogenetic record in Nemer- 
tines. On this theory of course the lateral cords of Nemertines 
do not correspond to the ventral cords of Annelids. The 
latter are represented by the general ventral plexus of 
Nemertines and by the pedal plexus or cords of Mollusca. 
These ventral nervous systems appear to bear relations to 
the dorso-lateral ring-nerve similar to those of the subum- 
brellar plexus of Medusae to the circumferential nerve-ring. 
It will be recognised from these remarks that the 
conclusions to which I have arrived present distinct points 
of agreement with those of Balfour (1, p. 37&) and Sedg- 
wick (21) on the same subject, although attained throughout 
by an independent series of inductions. With both these 
writers I agree in tracing back the Molluscan nervous 
system to a primitively annular type, such as might be 
expected to exist in a Ccelenterate ancestor. Balfour 
derives the whole Molluscan nervous system from a 
peripheral nerve-ring which followed the course of a hypo- 
thetical ciliated ring: ; Sedgwick derives it from a broad 
plexus surrounding an elongated blastopore, such as occurs 
in existing Actinians. Sedgwick's theory was practically 
an alternative to Balfour's, but I find myself able to give a 


partial acceptance to both these views. For the nervous 
system of Mollusca appears to me to consist of two parts, a 
circumferential ring and a peri-blastoporal plexus. The 
circumferential ring, which was primitively associated with 
a ciliated ring, is represented by the pleuro-visceral nervous 
system, which I have shown to possess significant relations 
with the velum or prototroch of the larva ; and the peri- 
blastoporal plexus seems to me to be recognisable in the 
pedal nervous system, which in primitive Molluscs has a 
very diffuse plexus-like arrangement, and in Amphineura, 
at any rate, reveals its peri-blastoporal character in the 
cerebro-pedal connectives in front and its connectives with 
the supra-rectal abdominal ganglion behind. 


(i) Balfour, F. M. Comparative Embryology, ii., 1885. 

(2) BOUTAN. Arch. Zool. Exp. (3), i., pp. 221-265, 1893. 

(3) Bouvier. Systeme Nerveux des Prosobranches. Ann. Sci. 

Nat. (7), iii., 1887. 
{id) Bouvier. Comptes Rendus, cxiv., p. 1281, 1892. 

(4) BOUVIER. Comptes Rendus, cxvi., pp. 68-70. 

(5) Duvernoy. Mem. sur le systeme nerveux des Mollusques 

Acephales. Mem. Acad. Sci. Paris, xxiv. 

(6) ERLANGER, R. von. Zur Entwicklung der Paludina vivipara, 

I. u. II. Theil. MorpJi. Jahrbuch, xvii., 1891. 

(7) ERLANGER, R. VON. Zur Entwicklung von Bithynia ten- 

taculata. Mitth. Zool. Stat. Neapel. x., 1892. 

(8) HALLER, BELA. Die Organisation der Chitonen der Adria. 

Arb. Zool. Inst. Wien, 1882-3. 

(9) HALLER, BeLA. Untersuchungen Liber marine Rhipido- 

glossen, I. Morph.Jahrb., ix., 1884. 

(10) Haller, Bela. Die Morphologie der Prosobranchier. Morph. 

Jahrb., xiv., 1888. 

(11) HALLER, BELA. Studien liber Docoglosse u. Rhipidoglosse 

Prosobranchier, 4I-0, 1894. 

(12) HUXLEY and PELSENEER. Report on Spirula. "Challenger" 

Reports, Zool, part lxxxiii., Appendix, 1895. 

(13) KERR, J. G. On some Points in the Anatomy of Nautilus 

pompilius. Proc. Zool. Soc, part iii., 1895. 

(14) Korschelt u. Heider. Lehrbuch der Entwicklungs 

geschichte, iii., 1893. 


(15) KOWALEVSKY, A. Embryogenie du Chiton polii. Ann. Mus. 

Hist. Nat. Marseille., Zool., i., 1883. 

(16) LANG, ARNOLD. Lehrbuch der Vergl. Anat, 3 heft. 

(17) Lankester, E. Ray. Mollusca. Encycl. Brit., ninth 


(18) PELSENEER. Contribution a l'etude des Lamellibranches. 

Arch, de Biol., xi., p. 166, pi. vi., fig. 3, 1891. 
(i8#) PELSENEER. Classification Generale des Mollusques. Bull. 
Sci. France et Belg., xxiv., p. 346, 1892. 

(19) PELSENEER. Introduction a l'etude des Mollusques. Bruxelles, 


(20) PLATE. Bemerk. lib. d. Phylogenie u. d. Entstehung d. 

Asymmetrie d. Mollusken. Spengel's Zool. Jahrbiicher, Abth. 
f. Anat. u. Ont., ix., i., p. 169, 1895. 

(21) Sedgwick, Adam. On the Origin of Metameric Segmenta- 

tion. Quart. Jour. Micr. Sci., xxiv., 1884. 

(22) SlMROTH. Bronn's Klassen u. Ordnungen des Thier-Reichs . 

Mollusca. I. Amphineura u. Scaphopoda, 1892-94. 

(23) Thiele, J. Ueber Sinnesorgane der Seitenlinie und das 

Nervensystem von Mollusken. Zeit.f. IViss. Zool, xlix., pp. 
385-432, 1890. 

(24) Thiele, J. Die Stammesverwandschaft der Mollusken. Jena 

Zeit., xxv., p. 480, 1 89 1. 

(25) Thiele, J. Beitrage zur Kenntniss der Mollusken. Zeit. J. 

Wiss. Zool, liii., p. 578, 1892. 

(26) WlLLEY, A. Natural Science, vi., p. 412, 1895. 

Walter Garstang. 


( Concluded. ) 

THE position of the glucosides in vegetable metabolism 
has been for a long time a subject of considerable 
controversy, which has, however, been most largely con- 
cerned with tannin. The details of its formation, its locali- 
sation and its fate have been discussed at great length, but 
the discussion has been largely conducted on the lines of 
hypothesis and analogy rather than experiment. The con- 
clusions reached by such a method of treatment have some- 
what hastily been applied to all glucosides, as if tannin were 
eminently the typical one. There are now reasons for 
thinking that so far from this being the case it is especially 

The number of oflucosides known has increased con- 
siderably in recent years as our investigations into plant 
metabolism have been pursued, and increasing knowledge 
of them forces the conviction more and more upon us that 
they take a more or less active share in the nutritive pro- 
cesses, possibly direct, but more probably through certain 
of the products to which they give rise on decomposition. 
They are not so markedly reserve stores for seeds as are 
many of the bodies we have already discussed, though many 
seeds, and notably many of those of plants of the Rosaceae 
and Cruciferse and orders allied to these, contain them in 
quantity together with other reserves. They occur, how- 
ever, in other parts of the plant, not quite as circulating 
reserves, but rather as transitory stores for more localised 
growth and nourishment. The old advocates of their 
nutritive functions rested their case largely on the presence 
of sugar in the glucoside molecule, and held that this is the 
body which is available for the constructive processes of the 
organism. There are, however, reasons for holding that 
this view is too limited a one. and that some of the other 
products of their decomposition may be as valuable as the 
sugar, if not of even greater importance. 


The glucosides that have attracted most attention during 
recent years are those which occur in the plants belonging to 
the families already mentioned, the Rosacese, the Cruciferae, 
and other orders which show affinities with these. These 
plants contain, very widely distributed through their tissues, 
amygdalin and sinigrine or myronate of potash respectively. 
Of these the former is perhaps the most interesting, as 
from its decomposition by enzyme agency there is produced 
hydrocyanic acid, which has always been regarded as most 
virulent in its action upon all living things. The existence 
of this noxious principle in the plant has perhaps been partly 
the cause of the readiness of botanists to class the glucoside 
which yields it, and hence the whole class of glucosides, 
among the products of excretion. 

The localisation of the amygdalin is calculated to throw a 
good deal of light upon the question of its probable function 
and fate. For many years attention has been given to it, 
at first, owing to imperfect methods of research, without 
much practical result. Improvement in technique has, 
however, yielded very valuable results, and has led to 
conclusions greatly at variance with those held thirty 
years ago. Thome (60), who wrote in 1865 upon the 
nutritive materials contained in the sweet and bitter al- 
monds respectively, said that amygdalin occurs in the 
parenchyma of the cotyledons of both varieties, and that 
its corresponding enzyme, emulsin, is only present in the 
bitter almond, being localised in the weak fibrovascular 
bundles that are in the cotyledons. This statement has 
been shown to be the exact converse of the truth. Portes 
(61), who worked twelve years later, showed that the gluco- 
side and the enzyme occupy different parts of the seed, the 
former being distributed in the cotyledonary parenchyma, 
while the latter is to be found in the axis of the embryo. 
Pfeffer (62), in his Pflanzenphysiologie, suggests that this 
localisation is not accurate, and that the two bodies probably 
occupy the same cells, the only degree of separation being 
that the ferment is in the protoplasm and the glucoside dis- 
solved in the cell-sap. In 1887 Johansen (63) by chemical 
methods succeeded in ascertaining the distribution of the 


two bodies in the seeds. He found the emulsin to be pre- 
sent in both varieties of the almond, and to be chiefly 
localised in the fibrovascular bundles. He further ascer- 
tained that the glucoside, amygdalin, is only present in the 
cotyledonary parenchyma of the bitter one. The absence 
of the glucoside from the seed of the sweet almond points, 
of course, to the conclusion that even if it be a nutritive 
body it is not one of very great prominence in the nutrition 
of the embryo on germination. 

Guignard has published within the past few years a 
series of researches which deal primarily with the localisa- 
tion of the enzymes which decompose the glucosides, but 
which incidentally throw a certain light upon the occurrence 
and meaning of the latter. In his first papers (64) he treats 
of the amygdalin which is found in the almond and in the 
cherry laurel, in the latter of which it is found to have a 
fairly copious distribution. He confirms Johansen as to its 
position in the seed of the almond, and still more closely 
localises the enzyme. In the laurel (Prunus lauro-cerasus) 
the parenchyma of the leaves as well as of the axis appears 
to contain it in solution in the cell-sap. The occurrence of 
the emulsin is confined to the neighbourhood of the con- 
ducting- tissues, it being chiefly found in the endodermis 
round the fibrovascular bundles. In the bundles of the axis 
of the embryo in the almond the ferment occurs in the many 
layered pericycle, chiefly outside the bast. The distribution 
of the amygdalin is not definitely known. It may happen 
that the fluid sap containing it may travel along the cellular 
tissue, and the occurrence of the ferment which decomposes 
it, in the immediate neighbourhood of the conducting tissues, 
suggests that it is charged with the duty of preparing from 
the glucoside certain nutritive products that may easily make 
their way to the conducting tissues, and so travel to the 
actual seats of constructive metabolism. That sugar so 
travels is of course a matter of every-day experience, but 
whether or no the remaining products are made use of in a 
similar way is open to discussion. On the other hand it 
may be that the amygdalin descends by the conducting 
tissue of the bast and undergoes decomposition as it passes 
downwards, yielding simpler products to the young cortex. 


In the face of the problem of the utilisation of the bodies 
resulting from the action of emulsin upon amygdalin great 
importance must be ascribed to the recent work published 
by Treub on the occurrence and meaning of hydrocyanic 
acid in the tissues of Pangium edule (65), one of the 
Bixacese. This compound, according to the author, does 
not occur as a glucoside, but in the free condition, and is 
present in relatively large amount. Greshoff found more 
than 1 per cent, to be hydrocyanic acid of the dry weight 
of the plant in one sample among many others analysed. 
A brief resumi, of the author's conclusions seems not to be 
out of place here, as throwing light upon the question of 
the nutritive value of the glucoside of the laurel. Indeed 
it seems not improbable that the hydrocyanic acid itself may 
be regarded as, in some cases at least, a reserve material. 

Treub has made a careful investigation into the 
localisation of this principle in the plant, using as his 
method the reaction given in the formation of Prussian 
blue when hydrocyanic acid comes in contact with a ferric 
salt in the presence of hydrochloric acid. The reaction is 
very distinct and takes place well in the interior of the 
cells, causing those which contain the hydrocyanic acid 
to stand out with great distinctness. 

In the whole of the adult axis, both stem, root and 
peduncles, he finds it to exist in quantity in the conducting 
tissue of the bast and pericycle. In the leaves it is still in the 
same regions, but is more widely spread, nearly all the 
parenchymatous tissue of the blade containing more or less 
of it. The epidermis especially is noteworthy, showing it 
present in the basal cells of the hairs which the leaves bear, 
and in certain idioblasts which contain also crystals of 
oxalate of lime. In the young fruits and those which are 
growing a considerable quantity is present, partly in the 
bast and partly in parenchyma outside the conducting tissue. 
In the seeds there is an accumulation in the peripheral 
layers of the endosperm and in other cells of the same 
tissue abutting on the embryo. 

In these regions, and in the cortex, and sometimes the 
pith of the axis, Treub describes the hydrocyanic acid as 


existing in special cells which are sharply marked off from 
the others round them when stained as above described. 
These special cells vary a good deal in number, apparently 
according to the amount of the acid present in the plant, 
and have no very specially regular distribution. Indeed it 
seems probable that any cell of the tissue may become a 
centre of deposition of the acid. Generally, if not quite 
isolated, they only occur two or three together. Certain of 
the fibres of the pericycle may be observed almost similarly 

Treub further says that these special cells of the cortex 
or of the pith derive their supply of hydrocyanic acid from the 
conducting tissue of the bast and that the amount of them 
and consequently of the acid varies with the condition of 
the stem. 

Tracing the hydrocyanic acid upwards through the axis 
by means of longitudinal sections it can be found to extend 
throughout its whole length, but to disappear at a little dis- 
tance from the growing point, the apical meristem of which 
contains none. 

It is impossible to avoid being struck with the similarity 
here exhibited to the fate of sugar, amides, etc., which as 
we have seen can be traced up to the seats of constructive 
metabolism and there cease, apparently giving rise to 
protoplasm. If this be so, the hydrocyanic acid must be 
regarded as a plastic material, unsuitable as at first sight it 
would appear for that purpose. 

This view is supported by several observations which 
the author details at some length. He finds that in the 
apices of young shoots which have suffered an arrest of 
growth, there are more of the special cells containing the 
hydrocyanic acid than there are in similar ones which are 
undergoing rapid elongation. That is, where there is active 
consumption of plastic material there is no accumulation of 
the acid, but where plastic substances are compelled to remain 
unused, hydrocyanic acid is one of such stored bodies. 

Another series of observations considerably strengthens 
this view, while it points more definitely to the ultimate 
purpose of the acid. In many of the special cells the latter 


may be seen to be accompanied by quantities of proteid 
substance. Taking young cells near the apex of the shoot 
the special cells contain the hydrocyanic acid alone, showing 
that it precedes proteid in the time of its occurrence. A 
little farther back the proteid can be detected, and gradually 
as sections are taken at increasing distances from the apex 
it increases in amount while the acid diminishes. As the 
active life of the cells becomes less and less vigorous, the 
proteid becomes more and more preponderating in the cell 
contents, and ultimately cells are found which contain 
proteid only, the hydrocyanic acid having all disappeared. 
The same succession of events can be seen if the develop- 
ment of the pericyclic fibres be traced towards the apex of 
the stem. 

There seems from these observations to be very strong 
reasons for supposing that hydrocyanic acid is a nutritive 
substance and leads at any rate in these plants to the 
formation of proteid. 

Treub holds that this is its immediate function ; he 
believes it to be primarily formed in the leaves, principally 
in the basal cells of the hairs and the idioblasts with calcic 
oxalate in the epidermis of the leaves. Thence it makes 
its way to the conducting tissues of the bast and pericycle 
and travels to the apical meristems. It is thus primarily a 
body originating only in the constructive processes, and not, 
as in the cases of the almond and cherry laurel, the product 
of a decomposition of a glucoside. Indeed Treub says very 
emphatically : " L'acide cyanhydrique du Pangium edule 
n'est pas un produit de decomposition ou de desassimila- 
tion," basing the statement on both indirect and direct 
arguments. The former are founded on the localisation of 
the product in the bast and pericycle and its evident trans- 
portation by the bast tissue. The latter involve the 
consideration of its localisation with a material which 
serves as a temporary proteid reserve in the same elements 
of the tissues, and the order of appearance and disappear- 
ance of the two substances in such special cells. 

That hydrocyanic acid can subserve not only the for- 
mation of temporary reserves of proteid but can be used, 



immediately after its first formation, by the leaves in which 
it is formed also appears certain. When plants whose 
leaves contain it are put for some days in the dark the acid 
gradually disappears, and as usual in such cases their whole 
metabolism suffers. On being again illuminated the vital 
processes gradually resume their activity. If a plant be 
put in the dark till nearly all the acid has gone from 
the leaves and then it be brought into the light, the little 
that remains is soon removed by the returning activity of 
the metabolism. 

That the acid is used, and not simply transported from 
the leaves, can be shown in another way, by cutting a 
circular section through the conducting tissue of the petioles, 
when removal by transport becomes impossible. Yet the 
hydrocyanic acid disappears gradually. 

It was said above that in some cases the hydrocyanic 
acid itself might be looked upon as a reserve material. 
This seems to be the case in the special cells described by 
Treub in the cortex of plants when they do not contain 
also proteid. In such cases we seem to have temporary 
reservoirs to supply local and transitory needs and to 
supplement the current passing along the bast. " Dans 
les endroits non on pas suffisamment desservir pour le 
systeme conducteur liberien ces usines locales prennent 
naissance, et en plus grand nombre, a mesure que la 
plante a on aura besoin dans ces endroits de plus de 
substances plastiques." Thus in the older part of the 
stem, where the active life is confined almost altogether 
to the cortex, the latter contains many of these special cells, 
while they are absent from the rest of the fundamental 
tissue. Where they are present, as in certain portions of 
the petioles, active life continues, although it may be de- 
cadent in other parts. 

This temporary storage comes out very prominently in 
the cases of the developing fruit and seed. At the base of 
the former, just above its point of junction with the pedicel, 
there is a very marked accumulation of the hydrocyanic acid, 
the cells staining blue under the treatment described being 
much more numerous than lower down the stalk. The 


peripheral layer of the seed in its young condition is 
also supplied very fully with these local reservoirs. We 
appear to have here a deposit laid down to supplement the 
regular stream which is passing all about the plant by means 
of the conducting tissue of the bast. It is doubtless derived 
from the circulating supply, for if the latter be interrupted 
by a section passing across the stem through its path, the 
disappearance of the acid takes place from the bast tissues 
below the wound some time before it does from the isolated 
special cells of the cortex. 

From the work of Treub and of Guignard then it seems 
increasingly probable that the glucosides are reserve 
materials, and not simply bye-products or products of 
excretion. Nor is it apparently only the sugar in them 
which has a nutritive value, but the other products of their 
decomposition have a particular part to play in the meta- 
bolism. This is certainly the case with hydrocyanic acid, 
and no doubt further investigation will show that it is the 
same with other products similarly formed. 

Guignard (66, 67) has made similar researches to those 
already described upon the plants of the natural orders 
Cruciferae, Capparidaceae, Tropceolacese, Limnanthaceae, 
Resedaceae and Papayaceae ; which all contain the ferment 
myrosin, a body capable of decomposing more than one 
glucoside. There are several of the latter compounds 
found in this group of plants, the best known of which are 
sinio-rine, anc j sinalbine. Siniorine is found in the black 
mustard (Brassica nigra), and is often called myronate of 
potassium. On decomposition it yields besides sugar a vola- 
tile body, sulphocyanate of Allyl, and potassic hydrogen 
sulphate. Sinalbine, as its name implies, is found in the 
white mustard (Sinapis or Brassica alba). When decom- 
posed the volatile constituent is found to be sulphocyanate 
of orthoxybenzyl. Others, the composition of which is not 
yet fully known, are those of the watercress {Nasturtium 
officinale) which yields phenyl propionic nitrile, the common 
cress {Lepidium sativum) affording the nitrile of alpha- 
toluic or phenylacetic acid. Though the fate of these 
complex volatile bodies has not been investigated, it is 


noteworthy that some of them at any rate contain cyanogen 
compounds, which may well be utilised after the manner of 
hydrocyanic acid itself as established by Treub. 

Their distribution in the plants appears to follow that of 
the amygdalin in the Rosaceous group, but very little 
definitely is known on this head. The enzyme which splits 
them up is according to Guignard always found in special 
cells which do not contain the glucoside. 

Very closely allied to the group of the glucosides is 
that of the tannins, about the importance of which there 
has been a good deal of controversy. Some of them are 
no doubt glucosides, yielding among their products of de- 
composition gallic acid and sugar. Others are apparently 
not so associated with a carbohydrate group. They are 
very widely distributed, and often occur not only in parts of 
plants which are devoted to storage of materials, but in the 
tissues where active metabolic work is going on. The 
task of deciding whether or no they serve as reserve 
materials or as bye-products is consequently not easy. 

The two views have been strenuously supported by 
different writers. Sachs, while working on the germination 
of the Scarlet-runner (68) in which tannin is comparatively 
plentiful, suggests an antithesis between carbohydrates 
and proteids on the one hand, and the tannins and colour- 
ing matters on the other, the latter being in his opinion 
only bye-products. He advances in support of his view the 
fact that they appear or increase with renewed growth of 
the embryo, instead of diminishing as reserve materials 
should do. Their appearance is coincident with the chemical 
changes in the undoubted reserves which lead to the 
utilisation of the latter. The same view is advanced by 
Schell (69), who suggests that in some cases, however, it 
may be a nutritive product. In the germination of certain 
oily seeds, chiefly of plants belonging to the Boroginaceae, 
tannin, which is present in addition to the oil, diminishes 
in quantity during the germination. In the stem of the 
mature plant there is during the winter a considerable 
quantity of tannin which almost vanishes as spring ad- 
vances. On the other hand he finds in certain almost parallel 
cases that the tannin accumulates instead of diminishing. 


The view that these bodies have a nutritive value has 
been supported with some emphasis by other writers. 
Wigand associated it very closely with the carbohydrates, 
and thought it was an essential factor in vegetable meta- 
bolism. Wiesner also supported the view of its carbohy- 
drate relationships, and indicated a probability that it stands 
between the starch and cellulose groups and the great class 
of resins, etc. The latter relationship has been again 
brought forward by Hillhouse (70), who found in Pinus 
sylvestris that as resin increases in the stem tannin dimin- 
ishes in like proportion, and that the cells surrounding the 
resin ducts invariably show its presence. Hartig suggests 
that tannin remains in the oak through the winter in the 
form of grains similar to starch grains, but distinguishable 
from the latter by characteristic reactions. These grains, 
he says, are dissolved and utilised in the spring. In his 
later writings Sachs inclines to the same view ; he says 
that besides those which must be looked upon as excreta or 
bye-products, some of the tannins of the oak are most likely 
to be regarded as reserve products, on account of their origin 
and disappearance and their behaviour generally during the 
growth of the plant (71). 

The localisation of tannin in the different parts of the 
plant does not give us much assistance in determining which 
of these views has most to support it. It is often found in 
special sacs in the midst of metabolic tissues ; it is very 
frequently found in epidermal cells, either in the interior or 
saturating the cell wall ; it is extremely prominent in bark. 
These positions certainly suggest that it is of but little value 
as a food-stuff; on the other hand it is often abundant in 
assimilating parenchyma in which starch formation is pro- 

In Hillhouse's paper (70) already alluded to, the author 
describes a considerable number of observations he made to 
determine whether or no a disappearance or diminution of 
tannin could be detected in the spring, and if so, whether it 
was a reasonable conclusion that such diminution indicated 
a utilisation of the vanished portion. 

He investigated a large number of trees in which tannin 


is present in greater or less amount, and noted the changes 
in the amount present in winter and in spring in their various 
tissues. He concludes that in no case is there noticeable a 
diminution of tannin in early winter as starch accumulates, 
and there is no sign that the starch is formed at the expense 
of the tannin. When growth recommences in the spring, 
instead of tannin disappearing from the older tissues it makes 
its appearance in quantity depending on the amount of 
growth. The tissues of the bud are commonly crowded 
with it. Hillhouse's experiments proceeded upon three 
lines. In the first place plants or parts of plants rich in 
tannin were made to grow under conditions in which assim- 
ilation of C0 2 was impossible ; a second set of experi- 
ments consisted of germinating in darkness seeds containing 
tannin ; and finally corms were investigated to see whether, 
as their nutritive material was transported to the newly- 
formed corm springing from them, tannin was transferred 
together with the starch. 

In no case was any diminution or transference found, 
except in the case of Pinus sylvestris already alluded to, 
when the probability of the tannin being an antecedent of 
the resin became evident. 

Those tannins which are undoubtedly glucosides must, 
however, be of some nutritive value, as they give off sugar 
on decomposition taking place. There is some evidence to 
show that during the ripening of certain fruits part of the 
sweetness is derived from an astringent principle resembling 
and probably identical with tannin, which diminishes in quan- 
tity as the fruit matures (72). 

A similar uncertainty as to its physiological meaning 
must for the present be associated with phloroglucin and 
the compounds into which it enters, which are to be re- 
garded as ethers corresponding to glucosides. There are 
two classes of these compounds, which have been described 
as phoroglucides and phloroglucosides respectively. The 
former include such bodies as hespentine, phloretine, etc., 
while the latter, which contain a sugar group in their for- 
mula, embrace aurantine, rhamnine, hesperidine, etc. They 
are somewhat difficult to localise, as the reactions they give 


are either not well ascertained or not particularly distinctive. 
The most reliable is perhaps that with vanilin in the pre- 
sence of hydrochloric acid. When this is made to react 
upon a cell which contains phloroglucin in the sap, the 
latter forms a fine precipitate of red granules which are 
composed of a compound of vanilin and phloroglucin, known 
as phloroglucivanilni. 

Phloroglucin appears to be often present in the plasma 
of meristem cells rather than in the vacuole, for when chlo- 
ride of vanilin is added to a tissue containing it the colouring 
mainly affects the protoplasm, some of the vacuoles remaining 
altosfether uncoloured. 

The distribution of phloroglucin, like that of tannin, 
leaves a good deal of uncertainty as to its physiological 
meaning. It has been investigated in recent years by 
Waage (73), who has carefully examined representative 
plants taken from almost all sections of the vegetable 
kingdom. Out of 185 plants submitted to experiment 
135 showed it to be present, but in very different quan- 
tities. Of the 135, 51 contained a very considerable 
quantity, 41 less but still a tolerably large amount, 
while in 43 though present only a feeble reaction could 
be obtained. Its distribution was to a certain extent 
regular, for the author states that if one species contains 
it, it is found with tolerable certainty in all the species of 
that genus. The plants of the Polypetalae as a rule show 
most, while the Gamopetalae and the Monocotyledons are 
on the whole poor in it ; lower down in the scale the Vas- 
cular Cryptogams and the Gymnosperms are charged with 
it to a degree intermediate between the other groups. 

Examining the tissues of such plants as contain a con- 
siderable quantity it may be found in meristems and in 
permanent tissues. In axial organs it occurs in the 
epidermis and later in the bark ; also in the parenchyma 
of the cortex, and in the sclerenchyma of the tissues more 
deeply seated. It is found sometimes in the endodermis ; 
also in the dead cell walls of the xylem parenchyma, fibres, 
and vessels. The medullary rays frequently contain a 
certain quantity. It is uniformly absent from the bast 


fibres and the sieve tubes, and may be present or not in 
the pith. When the epidermis contains it, it is usually 
in the hairs if any are present ; even root-hairs giving 
evidence of a certain amount. Taking the members of the 
axis, Waage found that roots as a rule contain more than 
stems, unless the latter be rhizomes, in which it is fairly 
abundant. Petioles and the peduncles of flowers contain 
less than branches. In plants where the axis is highly 
charged with it, there is generally a quantity also recog- 
nisable in the leaves, chiefly occurring there at the edges 
near the endings of the veins, and further in the neighbour- 
hood of the vessels of the latter. The palisade tissue of 
the leaf has usually more than the spongy mesophyll, and 
the upper has more than the lower epidermis. The seed 
as a rule contains but little, and that is only in the integu- 

If the disposition may be taken as any indication of 
its being a reserve material at all, the probability is that its 
value in the latter sense is but slight. The disposition of 
varying amounts in the medullary rays and its frequent 
presence in the cells of the cambium layer point possibly 
to its supplying nutritive material for the latter. On the 
other hand, its consistent absence from all parts of the seed 
except the integuments seems to indicate that storage of 
nutriment is not its main purpose. It may be that its value 
to the meristem tissues is based upon its easily oxidisable 
character, affording energy thereby, rather than being a 
reserve substance. Its occurrence in the leaves in the 
localities named suggests a formation in the mesophyll and 
a subsequent transport to the axial regions. But against 
the view of its value in metabolism as a reserve material 
we have the statement that light does not affect its forma- 
tion. It is in Waage's opinion found in the cell-sap as a 
general rule, rather than in either protoplasm or choro- 
plastids. It seems on the whole to be a product of 
destructive metabolism, for it occurs in the same cells as 
starch and sugar and may be derived from the latter by 
abstraction of three molecules of water, C 6 H I2 6 - 3 H 2 = 
C 6 H 6 Q 3 . It seems to resemble tannin in that it often 


increases with the greater development of the plant, and in 
being frequently plentiful in parts that are thrown off from 
the latter, such as old leaves, the coats of fruits, seeds, etc., 
and in regions withdrawn from active metabolism, such as 
bark and to a less degree epidermis. In a further paper 
Waage and Nickel suggest that it may possibly be a source 
of tannin, as the latter is generally found in the same 
parts as phloroglucin (74). Tannin does not appear, how- 
ever, to give rise to phloroglucin. 

Like tannin, therefore, phloroglucin appears to be on the 
whole an accessory product and only rarely to act as a 
reserve material. The compounds of it which contain 
sugar, i.e., the phloroglucosides, may serve as such, yielding 
sugar on their decomposition. 

In certain cases the alkaloids appear to serve as reserve 
materials, though their value in this direction is probably 
but slight. Many seeds which contain them in some con- 
siderable quantity lose them during germination, and other 
bodies, principally amides, replace them in the developing 
embryo or young seedling. This is especially the case with 
the seed of Lathyrus Sativus, an Indian species which 
contains sometimes as much as '5 per cent, of its dry 
weight of an alkaloidal product known as viciine (75). 

The possibility of alkaloids helping in such cases to 
form albuminoid materials or proteids has been pointed out 
by Jorissen (76) in his discussion of the chemical processes 
incident to germination, in which he claims for them a 
certain value as reserve materials. Heckel [jj) comes to 
the same conclusion. He carried out experiments with 
Sterculia acuminata, Strychnos Nux-vomica, Physostigma 
venenostim, and Datura Stramonium, and found in all these 
cases that during germination the greater part of their 
alkaloidal principles disappears. He claims that this 
disappearance is due to a transformation into assimilable 
substances under the influence of the embryo. If the 
latter be extracted from the seeds, and they be then sur- 
rounded by or buried in moist earth, the alkaloids remain 
for a considerable time unchanged. 

The conclusions of Jorissen and Heckel are disputed by 


Clautriau (78), who finds another explanation of the dis- 
appearance of the alkaloids during germination in a possible 
destruction of them as deleterious bodies which would 
affect prejudicially the development of the young seedling. 
He has ascertained with considerable precision the dis- 
tribution of the alkaloid in the seeds of Atropa Beliadona, 
Datura Stramonium, and Hyoscyamus Niger, and states 
that it is confined entirely to a layer of cells situated 
between the albumen and the integument of the seed, which 
when the latter is mature is very much reduced in its 
dimensions. This layer is much more prominent while the 
seed is ripening, consisting of many cells with very rich 
contents, the latter consisting of starch and albuminoid sub- 
stances as well as alkaloids. As the albumen grows, this 
nourishing layer gradually yields up both starch and pro- 
teids, while the alkaloid persists ; the cells become 
gradually nearly empty, and dry up considerably, ultimately 
becoming dead. In this condition they still contain the 
alkaloid, the quantity of which does not diminish during the 
changes described. When the seed is mature, this layer 
is very thin, the cells being flattened and compressed to- 
gether, forming a sort of membrane in which the alkaloids 
remain, partially or wholly combined with an organic acid. 

The nutritive value of the alkaloid seems improbable 
when we consider the disappearance from this layer of the 
starch and proteids, and the retention of the former. If it 
were then a reserve product it would in all probability ac- 
company the other undoubted nutritive bodies. Clautriau 
has obtained further information on this point by depriving 
seeds of Datura Stramonium of this alkaloidal layer and 
submitting them to germination, either in moist earth or in 
an atmosphere saturated with watery vapour. He found 
that under such conditions they germinated normally, and 
produced young seedlings which differed in no particulars 
from normal seedlings of Datura. 

Clautriau extended his researches to other plants than 
those named, particularly Conium maculatum, from which 
he obtained the same results. 

Examining the young seedlings grown under these 


conditions, no alkaloid being allowed to remain in the seed, 
Clautriau found that the active principle made its appear- 
ance in considerable quantity, and chiefly in thegrowingapices. 
The same thing was noticeable in the development of mor- 
phine in the poppy (79), where a more gradual formation 
was detected. Morphine does not show itself at the out- 
set of the development of the plant, but appears to be 
preceded by another alkaloid, giving very clear reactions, 
which does not seem to be identical with any of the nitro- 
genous principles extracted from opium. 

The conclusion that must be drawn from these investiga- 
tions is that these alkaloids, and hence probably all such bodies, 
are not to be regarded as reserve materials, but as bye- 
products or excreta, appearing coincidently with the active 
metabolic processes of the growing plant. 

Besides these accumulations of more or less complex 
organic compounds in the tissues of plants we meet with 
certain cases where inorganic material is deposited with a 
view to subsequent utilisation. These are, however, of 
much less importance and only occur in comparatively few 
plants. We have the well-known globoids in the aleurone 
grains of the castor-oil seeds, the seeds of Bertholletia 
excelsa and several others. From their disposition and 
fate, and from the fact that they afford a supply of phos- 
phorus, it is probable that we may include them in this 
group. In certain cases also the collections of crystals of 
calcium oxalate gradually disappear from the cells in which 
they are deposited, and so seem to minister to the needs of 
the plant for calcium, an element whose function, however, 
is still practically unknown. 


(60) Thome. Ueber das Verkommen des Amygdalins und des 

Emulsins in den bittern Mandeln. Botanische Zeitung, p. 
240, 1865. 

(61) PORTES. Recherches sur les amandes ameres. Journal de 

pharmacie et de chimie, t. xxvi., p. 410, 1877. 

(62) PFEFFER. Pflanzenpkysiologie, t. i., p. 307, 1881. 

(63) JOHANSEN. Sur la localisation de l'emulsine dans les aman- 

des. Ann. des. Sc. Nat. Bot., 7 ser., t. vi., p. 118, 1887. 


(64) GuiGNARD. Sur la localisation, dans les amandes et le 

Laurier-cerise, des principes qui fournissent l'acide cyan- 
hydrique. Journal de pharmacie et de chimie, 5 ser., t. xxi., 
pp. 233-289, 1890. 

(65) Treub. Sur la localisation, le transport, et le role de l'acide 

cyanhydrique dans le Pangium edule Reinw. Ann. du 
Jardin Botanique de Buitenzorg, xiii., p. 189, 1895. 

(66) GuiGNARD. Recherches sur la localisation des principes 

actifs des Cruciferes. Journal de Botanique (March), 1890. 

(67) GuiGNARD. Recherches sur la nature et la localisation des 

principes actifs chez les Capparidees, Tropeolees, Limnan- 
thees, Resedacees, et Papayacees. Journal de Botanique, 


(68) SACHS. Ueber die Keimung von Phaseolus multiflorus. 

Sits, der Wien Akad., 1859. 

(69) Schell. Physiologische Rolle der Gerbsaure. Bot. Jalires- 

berlcht, 1875. 

(70) HlLLHOUSE. Some Investigations into the Function of Tannin 

in the Vegetable Kingdom. Midland Naturalist, 1887-8. 

(71) SACHS. Vorlesungen iiber PJlanzenphysiologie, 1882. 

(72) BuiGNET. Ann. Chemie Phys., ser. iii., Bd. 61, 1861. 

(73) Waage. Ueber das Vorkommen und die Rolle des Phloro- 

glucins in der Pflanze. Ber. d. dent. bot. Gesell., November, 

(74) Waage and NICKEL. Zur Physiologie des Geitstoffs und der 

Trioxybenzol. Bot. Central., 1891. 

(75) RlTTHAUSEN. Journ. Jur pract. Chem., new series, vol. xxiv., 

p. 202, 1 88 1. 
(j6) JORISSEN. Les phenomenes chimiques de la germination. 
Memoires couronncs de I' Acad. Royale de Belgique, lxxxviii., 

P- 73- 

(77) HECKEL. Sur l'utilisation et les transformations de quelques 

alcaloi'des dans la graine pendant la germination. Comptes 
Rendus, January, 1891. 

(78) Clautriau. Localisation et signification des alcaloides dans 

quelques graines. Ann. de la Societe beige de Microscopie 
{Memoires), t. xviii., 1894. 

(79) CLAUTRIAU. Recherches microchimiques sur la localisation 

des alcaloi'des dans le Papaver somniferum. Mem. de la Soc. 
beige de Microscopie, t. xii. 

J. Reynolds Green. 

jj L I » & A R Yj 3u 



MANY parts of the interior of tropical Africa consist 
of wide grassy plains, occasionally varied by 
scattered trees, but usually very bare and monotonous in 
appearance. In the rainy season these steppes are green 
with vigorously growing grass, and patrolled by hundreds 
of antelopes and other kinds of game ; a few months after- 
wards when the rains are over, they are covered by 
blackened ashes and charcoal, and not a living creature will 
be visible except perhaps a few birds or a very occasional 

These fires are usually due to the natives, who find 
that the bush can be most easily cleared by their assistance, 
though they are often lighted to satisfy the childish delight 
in a big blaze which is characteristic of the Suahili porter. 

Their effects are most interesting, both economically 
and also in the way in which they entirely change the 
aspect of the vegetation. 

It is, of course, immediately obvious that all the valu- 
able feeding material of many square miles of luxuriant 
grass is by these fires entirely wasted ; but, besides this, 
the soil is never permitted to grow rich through the 
accumulation of leaf-mould and stems, and in fact the land 
is every year brought back into exactly the same condition. 
No true turf is formed, and the soil remains more like the 
subsoil in cultivated countries and never becomes in the 
least improved. 

The effect on the vegetation is very curious. The 
season of flowering for many trees and herbaceous 
plants is completely altered. A large number of low- 
growing herbaceous plants possess woody root-stocks or 
some sort of underground store of nourishment. With the 
very first shower of the rainy season, these stores send up 
flowering stems entirely without leaves, and the bare and 
blackened earth is studded with the bright purple flowers of 


Dolichos spp., the blue Pentanisia Schweinfurthii, little 
white Euphorbias, Lasiosiphon spp., etc. These all have 
the appearance of a flower cut off and planted in the earth, 
and give rise to remarks on the collector's carelessness in 
not bringing leaves when worked up by untravelled 
botanists. With the setting in of the rains, the stems 
begin to grow and produce leaves until, when the grass has 
sprung up, all these herbs are in full foliage. This habit is 
of great advantage to the flowers concerned, as insects 
can readily perceive the scattered flowers which in the grass 
would be quite inconspicuous. The same thing occurs in 
many of the trees. Several species of Dombeya, for example, 
send out their flowers at this foreshadowing- of the rains 
and are most conspicuous. 

Another curious effect of the fires is the manner in 
which trees are either kept down or obliged to protect 
themselves in some way against their action. In the more 
arid plains trees seldom exist, or if present occur in the 
form of stumps perhaps ten years old, but never able to 
grow higher than a foot or so. Such stumps put out every 
wet season vigorous shoots, which are annually burnt away 
and only the short stem with another layer of wood is left 
to survive. 

Of the trees which do manage to exist in spite of the 
annual conflagration, the most remarkable are the tree 
Euphorbias, often twenty to twenty-five feet high. These 
have angular fleshy branches protected by a leathery 
epidermis, and besides their milky juice, which contains 
gum, caoutchouc and other substances, have a large amount 
of mucilage or slimy matter in the ordinary tissue. This 
latter is a strongly waterholding substance, and the most 
violent fire seems unable to do more than scorch a very few 
of the outermost branches. 

It is a most curious fact that though when living they 
resist fires in this wonderful manner, dead branches make 
an excellent fire and blaze up most vigorously. I cannot 
understand this difference. 

Of the other trees which continue to thrive in these 
places, there are some seven species which grow in abun- 


dance ; there will be usually 500 of one of these species to 
every individual of some other kind. I brought home 
specimens of the bark of these six or seven forms, which 
were given to Professor Bretland Farmer for examination, 
who replied as follows : " I examined your specimens of 
bark and they all agree in possessing cells which show a 
certain amount of gummy degeneration of the cells in the 
bark, together with the presence of a considerable amount 
of sclerotic cells ; it seems not impossible that these two 
facts may be connected with the resistance of the plants to 
the fires, and I found as a matter of fact that, on comparing 
the rate of burning of these barks with that of laburnum, 
they were very slowly consumed. 

" I should have added that there are repeated periderms, 
and intermixed with the cork are the sclerotic cells 
already mentioned." Now the artificially produced cork 
of commerce shows great similarity in some respects to 
the cork of these fireproof trees. The process adopted 
both with the birch and the cork oak is to carefully peel off 
the cracked superficial layer of bark or " male cork " (this is 
known as "demasclage"). After this the layer of cork 
increases enormously and may perhaps attain to 17 cm. 
in thickness if left untouched : the result is the ordinary 
commercial article. I do not think that it is going too far 
to say that we have in grass fires a natural " demasclage " 
process, for they will certainly destroy the outer more or 
less dead tissues. 

From the researches of Henslow, 1 Tschirch 2 and 
Volkens 3 on desert plants, it may be considered proved 
that cutin, which most modern authorities consider nearly 
identical with suberin, is directly increased by dry and arid 
conditions, so that this direct effect is probably also of 
use in increasing the deposition of corky matter. Both 
evils — -the fire and the drought — have, as so often happens, 
brought about their own remedy. The sclerotic cells (or stone 
cork ?) may doubtfully be set down to the same cause, for 

1 Origin of Plant Structures. 

2 Angewandte Anatomie and Linnea, 1881. 

3 Flora der egypt. arab. IVuste. 


culture experiments (Duchartre and Henslow, loc. cit., p. 57) 
show that sclerenchyma may be directly diminished by a 
more moist atmosphere. 

The occurrence of gum is not so clearly dependent on 
the climatic conditions ; its use in these forms is, however, 
obvious enough, for all apertures by which water might be 
lost are, so to speak, gummed up. This is quite similar in 
physiological action to the drops of mucilage or gum which 
hermetically seal the vessels exposed by cutting across a 
branch of any ordinary deciduous tree. 

It is true that the production of gum is known to be 
most abundant in a dry and hot season, but according to 
the explanation given by Tschirch, loc. cit., p. 2 1 1 (and an 
identical account has been given me by Mr. Malcolm Dunn 
as the result of experience), this is due to the gum being 
squeezed out by the contraction of the bark following on a 
wet period, during which the masses of gum in the bark 
are greatly swollen. I cannot find any explanation of the 
actual cause of the change of cellulose into gum, but Mr. 
Malcolm Dunn states the general opinion that it is abundant 
after a severe shaking of the trees, as, for example, in a violent 
wind. Such places as those here treated of are certainly 
exposed to wind (otherwise they would be covered by 
forest, according, that is, to my experience), and it is possible 
that the wind may have assisted in starting gum formation ; 
but if, as is not unlikely, the wind acts indirectly by straining 
the layers of the cell walls, it seems more probable that the 
fierce heat of the fire, causing sudden and violent shrinking 
and warping of the bark, strains the cell walls in the same 
manner. This may of course be quite unproved, but the 
facts are sufficiently interesting to justify further research. 

G. F. Scott Elliot. 

Bcimce progress. 

No. 26. April, 1896. Vol. V. 



IF necessity be the surest prompter of invention, it is not 
too much to say that the necessity of the navigator 
has been a most potent factor in producing the observer of 
the elements of Terrestrial Magnetism. The traveller on 
land might rest during darkness until daylight enabled him 
to resume his journey ; but the seaman on the trackless 
ocean was dependent upon the indications of his compass 
by day and night ; and after the discovery of Columbus 
that the magnetic Declination or Variation of the needle 
from the direction of the geographical North varied in 
amount with the Latitude and Longitude, a new impetus 
was given to observation. 

The publication of Gilbert's grand discovery that the 
earth is a magnet and the director of the freely suspended 
needle, followed by the discovery of the secular change in 
the value of the Declination, naturally added to the desire 
of both landsmen and seamen to know as much as possible 
concerning that great magnet, both from purely scientific 
reasons and to meet the practical ends of the navigator. 
Thus the seventeenth and eighteenth centuries were re- 
markable for the number of observers both of the magnetic 
Dip and Declination. 

So important had a correct knowledge of the Declination 

become to the requirements of navigation, as early as the 

close of the seventeenth century, that Halley, under the 



immediate auspices of the Government, made his celebrated 
voyage over the Atlantic Oceans in a man-of-war, in order 
that intelligent observation should set at rest much that was 
doubtful. The results of this voyage, combined with the 
observations of previous navigators, were embodied in his 
celebrated chart of lines of equal value of magnetic Varia- 
tion or Declination, the first of its kind and of so convenient 
a form that charts of equal values of the three magnetic 
elements are to this day the most acceptable form for 
representing the combined results of magnetic observations 
made over large areas of sea and land, as well as of the 
special magnetic surveys which in recent years have been 
made in various countries. 

Here we may pause to consider the word Declination as 
applied to the angle which the direction of the horizontal 
magnetic needle makes with the true meridian. Many 
magneticians object to the word, but no better has yet been 
proposed or at any rate accepted ; the result being that 
while observers on land use the term, seamen adhere firmly 
to the expression "Variation of the Compass". This is as 
might be expected when it is remembered that navigators 
look upon the word Declination as connected with the posi- 
tion of the sun and other heavenly bodies, and would find it 
most inconvenient to have the same word in daily use, 
meaning two totally different things. 

During the eighteenth century charts of the magnetic 
Declination were published by Mountaine and Dodson, 
Bellin, and Churchman, and for their time may be con- 
sidered as fair approximations to the truth. Churchman's 
design was not only to give values of the Declination but 
to furnish the seaman with a means of ascertaining the 
Longitude, an ambitious project, especially as we now 
know there were probably considerable elements of error 
in these charts caused by local magnetic disturbance of the 
observing compass on land, and from the iron used in con- 
struction disturbing the compass on board the ships. 

This latter source of error was only beginning to be 
viewed in its true light at the close of the eighteenth 


In the years 1801-2 Commander Flinders of H.M.S. 
Investigator, then surveying the southern coasts of Australia, 
found that when his vessel's head was north or south by 
compass the observed Declination agreed very nearly, but 
when she lay with her head east or west, it differed largely. 
Moreover these errors on the east and west points of the 
compass had the opposite sign to those observed in Eng- 

Flinders, however, had supplemented the existing 
scanty knowledge of the distribution of the Dip over 
navigable waters by several observations of his own in 
northern and southern latitudes, and from these he drew 
the conclusion that the errors in the Declination observed 
on board ship were caused by magnetism induced by the 
earth in the vertical iron of the ship, and changed in value 
proportionally to change of Dip. Here Flinders was wrong, 
as the errors were really proportional to the tangent of the 

In spite of this mistake he was enabled from his know- 
ledge of the Dip to conceive the idea of so placing vertical 
bars of iron that they produced an equal and opposite effect 
on the compass to that of the ship in all latitudes, and thus 
invented what is now called the Flinders bar, one of the 
most important correctors of compass disturbance in the 
iron and steel ships of the present day. 

In 1 8 14 Flinders induced the Admiralty to have ex- 
periments made on board men-of-war at Portsmouth, 
Sheerness, and Devonport, to ascertain the amount of the 
magnetic disturbance of the compass caused by the iron in 
each ship. The chief reason for making these experiments 
was to show the necessity for ascertaining and applying 
these errors to ensure the safe navigation of the ships, but 
it had also a direct bearing in enabling observers to elimi- 
nate the hitherto inexplicable divergencies in the values of 
the Declination observed in different ships in the same 
geographical position. The results of these experiments 
bore no immediate fruit, for with the death of Flinders the 
subject was temporarily neglected. 

In 1 8 19, Hansteen published his Magnetismus der Erde 


with an atlas containing charts of the elements Declination 
and Dip for different epochs between the years 1600 and 
1787. These charts were in a large measure compiled from 
observations made with imperfect instruments and subject 
to the causes of error already mentioned attending both land 
and sea results. Hansteen, however, considered them of 
sufficient value to enable him to draw certain important con- 
clusions with regard to the cause of the secular change of 
the magnetic elements. Thus he not only concurred with 
Halley that the earth considered as a magnet had four 
poles or points of attraction, but computed their geo- 
graphical positions. Further than this, he computed that 
to account for the secular change these four supposed 
poles revolved round the terrestrial poles, each pole 
occupying a widely different number of years to complete 
the revolution. 

If these theoretical results had been true, a great 
advance would have been made not only in the science 
of terrestrial magnetism but in its practical bearing on the 
requirements of the present day. 

Although Humboldt had about the year 1800 shown 
that the intensity of the earth's magnetism varied with the 
latitude, the general distribution of that magnetic element 
was so little known that we may with our present extended 
knowledge consider that Hansteen's conclusions were based 
on insufficient data. In fact the idea of the earth being a 
magnet with four poles has long since been abandoned in 
favour of there being one pole with two foci of intensity in 
each hemisphere, and reasons will be given further on 
which tend to throw doubt on there being any revolution of 
these two magnetic poles round their adjacent terrestrial 

Subsequently to Hansteen's charts there appeared those 
of the Declination by Yeates, Duperrey, and by Barlow in 
1836. These were useful to navigation but helped very 
little towards the solution of the problem of the ever vari- 
able distribution of the earth's magnetism. 

Besides this by the year 1835 the iron-built ship had 
appeared on the ocean and a correct knowledge of the 


three magnetic elements became a necessity in solving the 
problems which the magnetism of different iron ships 

With Gauss's invention of the absolute horizontal force 
magnetometer in 1833, many hitherto unknown move- 
ments of the magnetic needle of the highest interest were 
discovered, which with the coarser instruments previously 
in use lay concealed. This discovery gave the desired 
impetus to the scientific men of that epoch, and the period 
included in the years 1835-45 was "a time of unparalleled 
activity in the extension of systematic and accurate mag- 
netical observations over the earth's surface ". 

Whilst most of the continental nations joined in this 
movement, the principal share in the work was divided 
between Germany, Russia, and England in Europe, and the 
United States in America. But before the splendid series 
of simultaneous observations made on the continent, and 
four British colonial observatories were organised, Gauss 
in 1839 published his general theory of Terrestrial Magnetism 
coupled with a series of charts of the three magnetic elements 
for the whole world, based upon observations made at 
ninety-two selected stations distributed over the earth's 
surface ; and it may be remarked that Gauss's charts not 
only gave results in fair accordance with observation in 
explored regions, but also as afterwards proved in Antarctic 
latitudes hitherto unvisited by man. 

The proof came in the years 1839-43, when Ross's 
Antarctic voyage of exploration was carried out in the 
interests of terrestrial magnetism. Besides the importance 
of a knowledge of the general distribution of the magnetic 
elements in those regions, one great aim of this expedition 
was to reach the south magnetic pole. This was found to 
be impossible, but sufficient data were collected to give its 
approximate position. Whilst this Antarctic magnetic 
survey was being completed, that of British North 
America was also undertaken with the result of the deter- 
mination of the locality of one of the foci of greatest 
intensity in the northern hemisphere. 

The results of these surveys formed, as will be well 


remembered, a valuable series of "contributions" to terres- 
trial magnetism by Sabine, and, coupled with every available 
observation between the years 1818 to 1876, formed the 
materials for the series of charts entitled " The Magnetic 
Survey of the Globe " for the epoch 1842*5. Each map 
gave normal lines of equal values of the Declination, In- 
clination and Intensity. Although it may be said that from 
the Arctic circle to the Antarctic, the direction of the lines 
was efficiently given by observation, the lines within those 
circles were largely taken from Gauss's computed lines 
modified to agree with observation. 

Another difficulty in compiling these charts of Sabine's 
with accuracy lay in the uncertain knowledge of the secular 
change then available, and which had to be applied to 
observations made so far apart in time. 

Sabine's charts are doubtless the best we have for the 
epoch 1842*5, but in the light of the requirements of 
modern science they leave much to be desired as regards 
the Antarctic regions. The observations south of 6o° S. 
were made entirely on board ships, where the errors of the 
compass sometimes exceeded 50° due to the horizontal 
forces in the ship, thus rendering accurate observations of 
the Declination very uncertain and correction of the observed 
Inclination very difficult ; besides which there are no 
records of the ship's disturbing force in the vertical direc- 

Naval requirements, however, did not permit of any 
delay in publishing magnetic charts affecting navigation, 
for in 1846 the Hydrographer of the Admiralty requested 
Sabine to provide charts of the Declination for the Atlantic 
Oceans from 6o° N. to 6o° S. These were largely used until 
Evans's chart of the Declination for the whole navigable 
world was issued in 1858. 

The excellent work of Flinders already referred to, of 
ascertaining from his knowledge of terrestrial magnetism 
the chief cause of the deviation of the compass in wood- 
built ships, and providing a corrector for those deviations, 
had to be followed up on a much larger scale and with 
more exact methods in the iron-built ship, which, in that 


period of activity in terrestrial magnetic science — 1835-45 — 
was rapidly increasing in numbers on the ocean. 

Thus in 1835 observations were made on board iron 
ships showing that they acted as a magnet on their com- 
passes, but there was nothing to show in the results what 
the causes of this condition of the iron ship were, until 
Poisson in 1838 published his celebrated "Memoir on the 
deviations of the compass produced by the iron in a ship ". 
This was a rigorous mathematical investigation of the 
subject, showing that the deviations of the compass were 
due to induction in the ship by the magnetic force of the 

If the iron ship had simply been built for service in one 
locality, a limited knowledge of terrestrial magnetism would 
have sufficed to elucidate the causes of her magnetic con- 
dition ; but she was destined to traverse every navigable 
sea over large changes of magnetic latitude, hence the 
necessity for an accurate knowledge of the distribution of 
magnetism over the great parent magnet, in order to 
determine the magnetic condition of her comparatively 
minute offspring the magnetised iron ship ; and this at all 
times and in all places in the interests of navigation. 
Observations of the terrestrial magnetic elements were 
therefore an absolute necessity if iron-built ships were to be 
substituted for those of wood. 

The ability to predict the deviation of the compass on 
change of latitude did not, however, satisfy Airy, for after 
a remarkable mathematical investigation of iron ship's 
magnetism of a less rigorous character than Poisson's, but 
sufficiently accurate for his purpose, he in 1839 proposed 
his methods of annulling the deviation of a ship's compass 
by means of magnets and soft iron, so arranged as to pro- 
duce equal and opposite magnetic effects to that of the 
ship. Provided with Airy's admirable and simple directions 
this method of correction was comparatively easy in one 
latitude, but experience at sea, especially in voyages to the 
Cape of Good Hope, showed that every iron ship required 
a different application of Airy's correctors. 

To discriminate between the amount that was to be 


corrected by permanent magnets, by horizontal soft iron, 
and by vertical soft iron, an accurate knowledge of the 
magnetic elements Dip and Intensity obtained from obser- 
vations on land and at sea was essential. 

Before dismissing the subject of the above application 
of magnetic observations, it may be remarked that we have 
now heavily armed, protected steel cruisers steaming over all 
parts of the world with less change of deviation of the 
compass than the wood-built Erebus and Terror of Ross's 
Antarctic expedition, and this remarkable result could not 
have been achieved if the terrestrial magnetic observer had 
not done his work. 

Moreover, if magnetic observations are not continued 
the secular change of the magnetic elements will soon 
commence to mar the precision with which our rapidly 
moving ships traverse the globe. 

The voyage of the Challenger in 1872-76 contributed 
the most valuable series of observations of the magnetic 
elements in modern times, when the large areas of the 
principal oceans traversed by that vessel during three and 
a half years are taken into consideration. These observa- 
tions, combined with those taken from every available 
source, both British and foreign, between the years 1865-87, 
formed the materials from which the magnetic charts of 
1880 were compiled (see vol. ii., Physics and Chemistry, 
part vi., Voyage of H.MS. "Challenger"). 

The Challenger only crossed the Antarctic circle at one 
point in longitude 78° E M and, therefore, although we know 
large secular changes to be going on south of 40 S. we have 
no measure of the amount, nor anything like an accurate 
knowledge of distribution of the earth's magnetism in those 
regions. This points to the necessity for a new Antarctic 

In the year 1888 the late Professor J. C. Adams was 
provided with a complete set of magnetic charts for the two 
epochs 1842-5 and 1880 previously mentioned, and as it 
was known he had directed his profound mathematical 
ability to the analysis of the results contained in them, it 
was hoped that some new and important light might be 


thrown upon the bare facts presented. His lamented death 
occurred without his publishing any results. 

If, however, reference be made to the report on the 
magnetical results of the Challenger, a discussion of the 
secular change is contributed founded in a great measure 
on a comparison of those charts. The outcome of this 
discussion is to throw considerable doubt upon the theory 
that the motion of the magnetic poles round the terrestrial 
is the cause of secular change ; in fact, that the magnetic 
poles remain fast, and we must look elsewhere for the cause 
whatever it may be. 

Magnetic observations have so far been considered in 
their all-important bearing as necessary to safe navigation 
in wood-built ships, and in a far higher sense as indispens- 
able to that of the iron- or steel -built ships which now 
cover the ocean ; the magnetic charts hitherto generally re- 
quired for these purposes being those on which normal 
lines of equal values have been given, but something more 
is now needed. 

Observation in comparatively recent years has shown 
that not only are there large " regional " magnetic dis- 
turbances extending over large areas of land, but that in 
moderate depths of water where the largest ship can navi- 
gate freely, the land below is also found to have considerable 
areas of local magnetic disturbance which, if not allowed for, 
may in thick or foggy weather lead ships into danger by 
seriously disturbing their compasses. 

The United States have done excellent work in pro- 
ducing charts of iso-magnetic lines, or charts in which the 
chief local magnetic disturbances are recognised, and the 
full results of observation recorded. The maonetic sur- 
veys of Riicker and Thorpe in the British Isles, of Moureau 
in France, of Rijckevorsel in Holland and elsewhere, have 
thrown considerable light on the magnetic conditions of 
those countries, but there remain whole continents to be 
covered by the observer. 

The direction of the iso-magnetics too from the deep 
sea to the dry land of the coasts is an extension of the 
subject, which the observer has hardly touched as yet, but 


one affecting the safety of navigation, as well as the question 
that has been raised, whether the water areas of the globe 
are as a whole more, or less magnetised than the land areas. 

To possess charts of iso-magnetic lines for even a few 
countries is an evidence of considerable advance in the 
knowledge of terrestrial magnetism, for if reference be 
made to Sabine's lines of intensity in his contribution on 
the magnetic survey of North-West America it will be 
found that he rejected certain observations he considered 
abnormal and defective, which Lefroy the observer con- 
sidered to be his best and naturally retained in his map ; 
the result being a considerable difference in the form of the 
curves adopted by the two magneticians, Sabine giving 
normal curves, Lefroy iso-magnetics. 

Respecting the local disturbances of the needle which 
have been so clearly proved, the question naturally arises, 
whence the cause of these disturbances ? It is now believed 
by many, if not finally accepted, that Rlicker and Thorpe 
have answered the question by the results of their laborious 
survey of the British Isles, coupled with Riicker's elegant 
investigations as to the permeability of specimens of the 
rocks taken from the localities in which magnetic dis- 
turbances were found. Their answer is to the effect that 
these disturbances, which have been found to extend over 
a region 230 miles long by about 110 miles broad, are 
due to induction by the earth's magnetism in rocks of dif- 
ferent permeability, either present as in the basalts on the 
surface or concealed by superficial deposits. 

These results are distinct from the extraordinary dis- 
turbances of the needle when in the immediate vicinity of 
permanently magnetised rocks, and when the radius of dis- 
turbance may be only as many feet as the extent of the 
regional disturbance is in miles. 

The points of interest in the question of regional 
magnetic disturbance are not confined to the magnetician, 
for the geologist cannot afford to neglect the valuable in- 
formation the magnetic needle affords. Thus although 
Rlicker and Thorpe have since made a second and more 
elaborate survey of the British Isles, their remark of 1890 


that "the kingdom can be divided into magnetic districts 
in which the relations between the direction of the disturb- 
ing forces and the main geological characteristics are so 
suggestive as to be worthy of careful statement and further 
investigation," not only holds good, but has received con- 

The mining engineer is deeply interested in a know- 
ledge of the Declination. Charts of normal lines are of 
great use to him whether above or below the earth's surface, 
but especially below when he has no other guide. To such 
an one a knowledge of regional magnetic disturbance as de- 
duced from surface observations is most important, as it 
tells him that he is in the neighbourhood of magnetic 
rocks, the disturbing effect of which on his compass needle 
may be far greater in the depths of his mine and turning it 
into a treacherous guide. 

We have now considered magnetic observations in a 
measure from the point of view of the immediate practical 
results which their scientific treatment produces, but who 
will say in this great maritime nation that the work of mag- 
netic observers, even if solely to make navigation poss'ble, 
is not worthy of the fullest consideration ? 

There is besides a vast field of inquiry for the observer of 
terrestrial magnetism in unravelling thesecretsof the earth con- 
sidered as a magnet, and the ceaseless change of its magnetic 
condition which the needle tells us of, for which no immediate 
practical result can be foreseen, yet is worthy of the attention 
of the ablest physicists and most advanced mathematicians. 

Inquiry into the causes of the secular change is one 
requiring the fullest attention, but observation has not yet 
done sufficient work. It certainly has done much in certain 
countries, and for a large portion of the world as regards 
secular change in the past, and data obtained for predicting 
future changes for a few years, but only one expedition has 
examined the Antarctic regions magnetically, and it is 
doubtful if any substantial progress will be made until a 
second expedition is made thither, one profiting by the 
experience of its precursor, and equipped with possibilities 
for work hardly hoped for by Ross. 


It may be remarked in passing that a remarkable 
alteration in the amount of the secular change has been 
noticed in the Declination and Inclination at the following 
observatories : Bombay, Batavia, and Hong Kong about 
the period of the eruption of Krakatoa in 1883. This may 
be only a coincidence, but may it not also point to the 
possibility that the changes below the surface of the earth 
which culminated in that mighty explosion, and may still 
be at work, have had, and continue to have, magnetic 
effects which are recorded by the needles at those observa- 
tories ? 

Critical investigations have for many years been directed 
to the elucidation of the causes of the observed diurnal 
variations of terrestrial magnetism. This work was long 
seriously retarded by the various methods adopted at different 
observatories for recording their results, obliging those who 
entered upon a comparison of such results to go through a 
tedious conversion of them into a common method. It may 
be said that the first large departure from this objectionable 
practice occurred when the International Polar Inquiry of 
1882-83 was undertaken by the various expeditions. 

This was an important step in the right direction, but 
there remains much to be done, as shown by the ten re- 
ports of the British Association Committee on "the best 
means of comparing and reducing magnetic observations ". 
Their last report consists of an able and suggestive paper 
by Dr. Chree, being the analysis of the results of five years' 
observations on "quiet days" at Kew, and is well worthy 
of attention as indicative of the present state of our know- 
ledge as regards the diurnal variation of the three magnetic 

Such investigations only encourage one in the hope that 
the much required observations in southern latitudes may 
be undertaken. The observatories at the Cape and Mel- 
bourne could do invaluable work if it were carried out on 
the lines of Kew, for example. 

Lastly, what more is there to be said about magnetic 
observations and their bearings ? We do not know why 
the earth is a magnet, the cause of the secular change of its 


magnetism, why it is subject to solar diurnal, lunar diurnal, 
sidereal diurnal and the other variations and disturbances, 
nor the cause of magnetic storms, although we can observe 
connections between them, earth currents, and aurorae. 
Whether the causes of all these exist below the surface 
of, or are external to, the earth, or are a combination of 
the two, has still to be learnt, and it seems hardly too much 
to hope that the restless needle will sooner or later be the 
means of opening up sources of knowledge invaluable to 
cosmical science, as well as to those only concerned with the 
planet upon which they dwell. 

When the causes of the secular change are understood 
there will be no difficulty in providing the navigator with 
magnetic charts for years in advance, much as the tides can 
now be tabulated for his use. In the latter case observa- 
tion has done its work for several frequented ports, in the 
former case a vast amount remains to be done, and the 
word that goes forth is still, as Lord Kelvin thrice said 
on a kindred subject connected with ships' magnetism, 
" Observe". 

Ettrick W. Creak. 



A FEW years ago a discussion of the cell-theory would 
have seemed superfluous. To-day, partly because 
of criticisms which have been directed against the theory, 
partly because of the great increase of our knowledge re- 
specting cell-structure, the advantage and even the necessity 
of such a discussion will be admitted by everybody who has 
read and reflected on the subject. In what follows, I 
propose to examine the cell-theory in the light of recent 
criticisms and researches. I set out with the intention of 
avoiding anything in the shape of polemical writing, but I 
fear that I have in places fallen away considerably from the 
course which I had proposed. In a much disputed subject 
controversv is inevitable, a circumstance which need not be 
regretted, for controversy is the whetstone of argument, and 
obliges those who engage in it to be doubly careful both of 
their facts and of the language in which they express them. 
My antagonists will, I hope, give me the credit of the 
desire to deal fairly with their arguments and criticisms, and 
will acquit me of unnecessary bitterness. It has been my 
object to elucidate the subject in hand rather than to try to 
gain a dialectical advantage. 

It is advisable, before entering on the examination, to 
have a clear conception of what the cell-theory really is. 
This is the more necessary because one of its most recent 
critics, Mr. Adam Sedgwick, has complained than nobody 
will define the theory in an exact manner ; it is, he says, a 
kind of phantom which takes different forms in different 
men's eyes. I have shown in another place that this state- 
ment is hardly fair, because there are some authors whose 
researches on cytology entitle them to speak with authority 
who have recently defined the cell-theory in a very precise 
manner, though it may be conceded that there are biologists 


whose views are not so exact, and who habitually commit 
themselves to statements which on careful examination may 
prove to be altogether untenable. 

It was pointed out some time since by Whitman, 1 and I 
have since emphasised the fact, 2 that in his broad generalisa- 
tions Schwann defined the cell-theory in a very exact manner, 
and that the words originally used by him are perfectly 
applicable to the cell-theory as it has been held up to the 
present time. In saying this, I do not forget that Schwann 
held some very erroneous views as to the nature and 
structure of cells, which he regarded as vesicles, filled with 
fluid, which made their appearance in a structureless matrix, 
named for this reason, a cytoblastema. But Schwann's 
work consisted of two parts, a statement of observations, 
which have proved to be entirely erroneous, and a theory 
of organisation, which has been very fruitful of results. He 
was careful to say that his theory was only a provisional 
explanation which suited the facts as nearly as possible, and 
it is a great merit of the theory that it afforded such an in- 
sight into organisation that the essential part of it did not 
cease to be serviceable long after the "facts" on which it 
was founded were shown to be, for the most part, false. 
We need not therefore concern ourselves with the fact that 
Schwann's conceptions of the origin and structure of cells 
were false, but we may examine his theory and see how 
much of it we may hold to, and how much we must reject 
at the present day. 3 

Schwann was a very cautious writer, and the quotations 
which are given below will dispose effectually of the state- 

1 C. O. Whitman, "On the Inadequacy of the Cell-theory of Develop- 
ment, "Journal of Morphology, viii., p. 639, 1893. 

2 G. C. Bourne, "A Criticism of the Cell-theory," Quart. Jour. Micro- 
scopical Science, xxxviii., p. 137, 1895. 

3 A large part of Schwann's theory of cells, viz., that part of it which 
compared cell-formation to the process of crystallisation, was soon shown 
to be untenable. But as this part was based on his erroneous views on 
the structure and origin of cells, I have passed it over, since the falsity of 
his views on this subject involved the falsity of as much of his theory as 
was founded on them. 


ment which stands in the first paragraph of Whitman's 
work, that he believed that in cell-formation lies the whole 
secret of organic development. There are, says Schwann, 
two possible theories on the subject of organic development: 
(i) The organism theory, namely, that there is an inherent 
power modelling the body in accordance with a predominant 
idea. (2) The physical theory, namely, that the funda- 
mental powers of organised bodies agree essentially with 
those of inorganic nature. Rejecting the former of these 
two theories as being outside the domain of physical science, 
Schwann went on to write : x " We set out with the sup- 
position that an organised body is not produced by a 
fundamental power which is guided in its operation by a 
definite idea, but is developed according to the blind laws 
of necessity by powers which, like those of inorganic 
nature, are established by the very existence of matter. 
As the elementary materials of organic nature are not dif- 
ferent from those of the inorganic kingdom, the source of 
the organic phenomena can only reside in another com- 
bination of these materials, whether it be in a peculiar 
mode of union of the elementary atoms to form atoms of 
the second order, or in the arrangement of these con- 
glomerate molecules when forming either the separate 
morphological elementary parts of organisms, or the entire 
organism. We have here to do with the latter question 
solely, whether the cause of organic phenomena lies in the 
whole organism or in its separate elementary parts. If 
this question can be answered a further inquiry still re- 
mains as to whether the organism or its separate elementary 
parts possess this power through the peculiar mode of 
combination of the conglomerate molecules or through the 
mode in which the elementary atoms are united into con- 
glomerate molecules." 

Is it not perfectly clear from this that Schwann fully 
recognised that there was a further question underlying 

x Th. Schwann, Microscopical Researches into the Accordance in the 
Structure and Growth of Animals and Plants. Translated by Henry 
Smith. London: Printed for the Sydenham Society, 1847. 


the cell-theory, and do not the words which he used with 
regard to the union of elementary atoms to form atoms of 
the second order show a prescience of the assumptions 
which would have to be made to explain the powers mani- 
fested by cells ? Because he confined himself to the one 
question, it is not fair to say that Schwann had not a clear 
appreciation of the importance of the other. I may relate, 
in this connection, an anecdote which will finally clear 
Schwann's reputation from the reproach fastened upon it. 
Professor Lankester tells me that about the time when a 
sort of jubilee was held in Schwann's honour at Liege in 
1878, he was introduced to him, and ventured to ask in 
the course of conversation how it was that after the publica- 
tion of his famous essay he had so long been silent. 
Schwann answered that he had not been idle, but that 
ever since he had been unsuccessfully occupied in trying to 
find out the secret of the constitution of the cell. 

To return to the question propounded by Schwann, 
does the cause of organic phenomena lie in the organism or 
in its separate elementary parts, the cells ? He answers 
very decidedly, in the separate elementary parts, and gives 
the following reasons for his answer: " All organised 
bodies are composed of essentially similar parts, namely, of 
cells ; these cells are formed and grow in accordance with 
essentially similar laws, and therefore these processes must 
in every instance be produced by the same powers. Now 
if we find that some of these elementary parts not differing 
from the others are capable of separating themselves from 
the organism and pursuing an independent growth, we may 
thence conclude that each of the other elementary parts — 
each cell — is already possessed of the power to take up fresh 
molecules and grow, and that therefore each elementary 
part possesses a power of its own, an independent life, by 
means of which it would be enabled to develop itself in- 
dependently if the relations which it bore to external parts 
were but similar to those in which it stands in the organism. 
The ova of animals afford us examples of such independent 
cells apart from the organism." 

A little further on he says : " In inferior plants any given 



cell may be separated from the plant and can grow alone. 
So that here are whole plants consisting of cells which can 
be positively proved to have independent vitality. Now as 
all cells grow according to the same laws, and consequently 
the cause of growth cannot in one case lie in the cell, and 
in another in the whole organism, and since it may be 
further proved that some cells, which do not differ from the 
rest in their mode of growth, are developed independently, 
we must ascribe to all cells an independent vitality, that is 
such combinations of molecules as occur in any single cell 
are capable of setting free the power by which it is enabled 
to take up fresh molecules. The cause of nutrition and 
growth resides not in the organism itself but in its separate 
elementary parts. . . . The manifestation of the power 
which resides in the cell depends upon conditions to which 
it is subject only when in connection with the whole or- 

The whole theory is very succinctly summed up in the 
following passage : " The elementary parts of all tissues are 
formed of cells in an analogous though very diversified 
manner, so that it may be asserted that there is one uni- 
versal principle of development for the elementary parts of 
organisms, however different, and that this principle is the 
formation of cells ". 

No doubt objection may be taken to-day to the uni- 
versality of this statement, but if we modify the last part 
of it and read " that the most general principle of develop- 
ment for organisms, however different, is the formation of 
cells," we shall have very nearly expressed the truth, as we 
know it at the present day. 

I have found it necessary to quote Schwann's work at 
considerable length, and to repeat more emphatically what 
I stated in my previous essay on Epigenesis and Evolution. 1 
Dr. Whitman, 2 in a reply which deals partly with my 

1 G. C. Bourne, "Epigenesis and Evolution," " Science Progress," 
vol. i., 1894. 

2 C O. Whitman, Evolution and Epigenesis. Boston : Ginn & Co., 


criticisms, and partly with the much more weighty argu- 
ments brought forward at the same time by Dr. Oscar 
Hertwig, says that my criticisms, in so far as they are 
directed against the inadequacy of the cell-theory of develop- 
ment, are largely the result of misunderstanding ; this may 
in part be true, but I cannot have misunderstood the simple 
meaning of his first paragraph, and I wish to insist on the 
fact that the cell-theory, as it was promulgated by Schwann, 
did not regard cell-formation as the whole secret of organic 
development, and that the cell was not, in the mind of the 
author of the cell-theory, the alpha and omega of both 
morphological and physiological research in the animal 
kingdom. If this is clearly understood at the outset, it will 
help to remove much possible misunderstanding. 

But, as Mr. Sedgwick has rightly said, we have to deal 
not only with what its authors thought, but with the cell- 
theory as it is understood and taught at the present day. 
I have already pointed out * that the most recent definition 
of the cell-theory is, to all intents and purposes, identical 
with the broader generalisations of Schwann. Dr. Oscar 
Hertwig writes : 2 " Animals and plants, so dissimilar in their 
outward appearances, are similar in the essentials of their 
anatomical structure, since both are composed of similar 
elementary parts which for the most part are only recognis- 
able by the microscope. . . . Since the cells, into which 
the anatomist resolves vegetable and animal organisms, are 
the bearers of the vital processes, they are, as Virchow has 
expressed himself, the vital units. Viewed from this stand- 
point the whole life process of a composite organism appears 
to be nothing else than the extremely complicated result of 
the individual life processes of its numerous and variously 
functional cells." This is simply a restatement in other 
words of two of the fundamental principles of Schwann, 
namely (1) that the elementary parts of all tissues are formed 

1 G. C. Bourne, "A Criticism of the Cell-theory," Quart. Jour. Micr. 
Science, vol. xxxviii., p. 137, 1895. 

2 O. Hertwig, Die Zelle utid die Gewebe. Berlin : R. Friedlander 
und Sohn, 1893. 


of cells ; (2) that the cause of nutrition and growth resides 
not in the organism but in its separate elementary parts. 

The attacks which have recently been directed against 
the cell-theory may be resolved into contradictions of these 
two fundamental propositions. On the one hand there is 
the considerable number of cytologists, whose opinions 
may be taken to be summed up in Whitman's essay on 
the inadequacy of the cellular theory, who deny the second 
proposition, and in so doing implicitly deny the truth of 
the first. They would say that the cause of nutrition and 
growth does not reside in the cells considered as elementary 
parts, but in parts still more elementary, the ultimate vital 
units of which the cells themselves are composed. On 
the other hand Mr. Adam Sedgwick denies the first pro- 
position in toto. He states boldly that the elementary 
parts of tissues are not formed of cells, but of a continuous 
mass of vacuolated protoplasm containing nuclei. 1 These 
objections, though they are raised from different stand- 
points, are not irreconcilable, but it will be convenient to 
deal with them separately. First let us consider the 
objections to the cell as an ultimate vital unit. 

These objections are of long standing. They were first 
brought forward by Briicke 2 in 1861 ; 3 not long afterwards 

1 Since this was written Mr. Sedgwick has published a further account 
of his views, which makes it necessary to modify this statement. See 

2 Ernst Briicke, " Die Elementarorganismen," Sitzungsberichte der K. 
Akademie der Wissenschaften, Wien, bd., xliii., p. 381, 1861. 

3 Delage points out that the merit of regarding protoplasm as an 
organised substance belongs to Dujardin, and not to Brucke. At the 
same time he points out the essential difference between Briicke's concept 
of organisation and that of Dujardin, greatly to the advantage of the latter : 
" La difference entre Dujardin et Brucke est tres simple. Le premier a 
devine l'existence de structures que le microscope demontre aujourd'hui ; 
tandis qu'en introduisant dans la conception de protoplasma cette notion 
acceptee avec enthousiasme, d'organismes tres compliques et invisibles, 
Brucke a ouvert la porte aux nombreuses theories speculatives qui cher- 
chent a imaginer la structure de ces organismes pour expliquer par elles les 
phenomenes de la vie." Delage adopts the expression organisation, saying : 
" Le protoplasme n'est pas simplement, comme on l'a cru longtemps, une 
substance chimique organique, mais il est organise, c'est-a-dire possede une 


Herbert Spencer followed with his theory of physiological 
units. Darwin's theory of pangenesis expressed the same 
idea, and more recently Nageli, De Vries, Wiesner, Weis- 
mann and others have entered the same or at least similar 
objections to the cell-theory. Even Oscar Hertwig, although 
he appears in the sentence above quoted to give his 
adherence to the view that the cell is a vital unit, abandons 
this concept, for all practical purposes, in the latter part of 
his book ; for he says, in a most unmistakable manner, 
that the cell is an organism composed of ultimate units 
which he calls idiosomes. 

Each author whose name I have quoted has a somewhat 
different account to give of the ultimate constitution of the 
cell. But the points on which they disagree are of subor- 
dinate importance ; they are all agreed on the main issue, 
that the vital activities manifested by the cell are not to be 
explained by the visible constitution and structure of the 
cell itself, nor by the mere chemical elements of which the 
protoplasm of the cell is composed. Each of them avers 
that the cell is organised, which means that it is made up of 
countless organic units of a lower order, differing among 
themselves, and arranged in groups and sub-groups within 
the cell in a manner analogous to that in which the cells 
themselves are arranged in a composite organism. Since 
there is so general an agreement in fundamental principle, 
1 am spared the necessity of examining each separate theory 
of ultimate vital units in detail ; should anybody wish for a 
condensed account of the various theories he will find it in 
Weismann's introduction to his work on the Germ Plasm. 1 

structure d'un ordre plus eleve que la structure atomique des molecules 
chimiques des composes organiques non vivants ". No fault can be found 
with this definition, but would it not be better to adopt some other term to 
express this extra complexity of structure rather than " organisation," which is 
inseparably connected with our ideas of the composition of the bodies of 
higher animals and plants? For Brucke the organisation of protoplasm 
was the same in kind as the organisation of higher animals : for Dujardin 
it was something different, and had best be expressed by a different term. 
Delage puts the word structure in italics. 

1 Still better in Delage's book, referred to further on. 


The point for present consideration is this : Is it neces- 
sary for the explanation of vital phenomena to assume 
the existence of ultimate vital particles, so minute as 
to be invisible with the best microscopical powers 
which we possess, each of which is to be considered 
as being in posse an independent organism capable of 
displaying some of the most characteristic of vital pheno- 
mena, viz., assimilation, growth, metabolism, reproduction 
by division ? 

As it will be necessary to refer frequently to these 
assumed vital units, I must call them by some name, and I 
shall use Weismann's term biophor, without meaning to 
exclude the conceptions of other authors, the pangenes of 
De Vries, the plasomes of Wiesner and so forth. I use the 
term biophor in the sense of Lebenstrager, the bearer of the 
vital properties, without necessarily implying that it pos- 
sesses all the particular properties assigned by Weismann 
to his biophors. 

Whatever the point from which the different authors 
have started, they all postulate the existence of minute 
biophors on the grounds that the vital phenomena exhibited 
by cells, say by an Amoeba, or by the ovum of a Metazoon, 
imply the existence of an organisation adequate to the 
production of the observed processes. The life processes 
are too various and too complicated in their kind to be 
explained by the visible constitution of protoplasm, even if 
it be allowed, as it generally is allowed, that protoplasm is 
not a chemical compound of fixed molecular composition 
but a mixture of many chemical substances, each having a 
molecular composition of some considerable complexity. I 
have already shown that Schwann himself was possessed of 
such an idea, which he expressed sufficiently clearly when 
he referred to " a peculiar mode of union of the atoms 
to form atoms of the second order," but he did not attempt 
to follow out the idea, confining himself to the inquiry 
into " the arrangement of these conglomerate molecules 
when forming either the separate elementary parts of 
organisms or the entire organism ". The term consdomer- 
ate molecule is in fact synonymous with the term biophor 


in the sense in which I am using it, for the biophor or 
ultimate vital unit is held to be an aggregate of chemical 
molecules ; the constitution attributed to it is that it is made 
up of many different kinds of molecules, just as a molecule 
may be composed of several different kinds of atoms. I shall 
have to refer aram to the difficulties which still remain if 
we accept the hypothesis that a group of different molecules 
is able to exhibit the vital functions which are necessarily 
attributed to a biophor. Before proceeding to criticism I 
must try to give as fairly as I can the grounds for believing 
in the existence of biophors. To put the matter as briefly 
as possible, the theories of ultimate vital units are the resuh 
of attempts to make a mental analysis of living substance 
Chemical analysis is impossible, for in the process the 
living substance is destroyed as such and becomes dead 
substance, possessed of different and much less important 
properties. One fact of great importance, however, is 
learnt from chemical analysis, and it was appreciated by 
Schwann, namely, that, to use his original words, "the 
elementary materials of organic nature are not different 
from those of the inorganic kingdom " ; hence it has been 
inferred, with all reason, that the powers of organic nature 
are essentially the same as those of inorganic nature, and 
are established by the very existence of matter. It is only 
necessary to mention this because there has recently been 
a tendency in some quarters to call in the assistance of some 
mysterious " vital force " ; a tendency begotten no doubt by 
the apparent futility of all attempts to find an explanation 
on mechanical and chemical principles of the fundamental 
powers of organic nature, assimilation and metabolism. 
This, however, need not detain us ; we have to consider 
the process of reasoning which, in default of assistance 
from chemical analysis, has led so many distinguished 
observers and thinkers to analyse the cell into other com- 
ponents, and those again into others of a lower grade, 
until the biophor, the smallest particle of possible life, is 

The weightiest reason which I have been able to dis- 


cover is given by von Sachs. 1 According to this author, 
whose views are in agreement with those of Nageli on this 
subject, it is necessary for the explanation of certain pheno- 
mena exhibited by organic substances that we should assume 
the existence of combinations of molecules which form very 
large numbers of small particles or micellae as Nageli calls 
them. One of the most important of these phenomena is 
the imbibition of water. Dry organic substances, such as 
gelatine, when placed in water, imbibe it and increase in 
volume to a very considerable extent. The increase in 
volume produced by the swelling up in water is almost 
equal to the volume of water which has been absorbed. 
The imbibition of water in such a case is something very 
different from the imbibition of water by a porous inorganic 
body, such as gypsum, unglazed porcelain, etc. The latter 
substances are full of small visible and invisible cavities or 
pores, which in the dry state contain air. The water passes 
into these cavities or pores according- to the laws of capil- 
larity, and in so doing displaces the air, which is forcibly 
expelled and can be collected and measured ; there is no 
pushing asunder of solid parts, as is shown by the fact that 
the porous body is not perceptibly enlarged by the water 
which has penetrated into it. But the water penetrating 
into gelatine expels no air, it does not enter by capillarity 
into spaces previously existent, but forces its way between 
the particles of the dry substance, pushing these asunder, 
and so causing the considerable increase in volume. The 
particles thus pushed asunder are the micellae, and although 
they are pushed further apart from one another, they do 
not completely lose their connection. Each micella may be 
regarded as being surrounded by an envelope of water 
when in the moist state ; in the dry state the micellae com- 
posing the substance are in mutual contact. This familiar 
phenomenon of the swelling of organic substances by the 
imbibition of water is contrasted by von Sachs with the 
process of solution of a salt. In the latter case the water 

X J. von Sachs, Lectures on the Physiology of Plants, translated by 
H. Marshall Ward. Oxford, Clarendon Press, pp. 205 and sq., 1887. 


seizes on the molecules of a crystal and takes them in 
between its own molecules ; in the former case the dry 
organic body seizes on the molecules of water and forces 
them between its own. These reasons are held by von 
Sachs and Nageli to be among the weightiest for regarding 
protoplasm as an "organised" body, in the sense of being 
made up of micellae, and not as being a structureless slime 
or fluid. 

No doubt they are weighty reasons for regarding or- 
ganic substances such as gelatine, starch grains, cell walls, 
etc., as being composed of combinations of polyatomic 
molecules into groups of a higher order, and there is no 
objection to giving these groups a name, such as micellae. 
But the admission that such groups exist does not really 
bring us much nearer to an explanation of the phenomena 
of life. Von Sachs himself points out that even in the 
region of pure chemistry it is necessary to assume that 
polyatomic molecules are grouped into closer molecular 
unions, thus giving rise to chemical properties which did 
not belong to the individual molecules. 

Gelatine, starch grains and cellulose are not living but 
dead substances, and the fact that the behaviour of dead 
organic substance finds an explanation on a theory of 
micellar structure is but a very small step towards the 
explanation of the very different behaviour of living sub- 
stance. The micellae may exist in the organic substances 
in question, but they are not to be confounded with biophors ; 
the very fact that the properties of dead substances may be 
attributed to their existence shows that they cannot be con- 
sidered as bearers of vital properties. 

In point of fact the living substance, which we generalise 
under the name of protoplasm, behaves quite differently 
in respect of the imbibition of water to the dead substances 
which are derived from it. An amoeba or an infusorian, 
living in the water, does not imbibe it as a mass of gelatine 
would. But when it dies in the same water it immediately 
begins to swell up, and eventually bursts and disintegrates. 
So that we see that with respect to this very property which 
is held to be a reason for assuming a micellar structure of 


protoplasm, the actual living substance does not exhibit the 
property, whilst the same substance when dead does. 
Clearly then, the admission that protoplasm has a micellar 
structure, that is, that it is composed of minute and invisible 
particles consisting of groups of polyatomic molecules, does 
not involve the admission that there are ultimate vital units, 
biophors, which reside in the cell-like organisms within the 
cell-organism. This distinction indeed has already been 
made and dwelt upon at some length by Weismann (op. 
cit., pp. 41 and 42). 

It follows then that whilst we may freely admit that 
protoplasm, and also various inert organic substances, are 
composed of micellae, and are therefore "organised" in the 
sense spoken of by von Sachs, we have still to consider the 
evidence for the existence of biophors. At the outset of 
this inquiry we meet with a difficulty in that the existence 
of biophors is assumed by most authors as a means of ex- 
plaining the phenomena of heredity, and this opens up a 
wide range of questions into which it is not the purpose of 
this essay to enter. But it has well been pointed out by 
Wiesner that if minute vital elements occur at all, those 
same units which make life possible, and control assimila- 
tion and growth, must also be the agents in bringing about 
the phenomena of heredity. This view, which commends 
itself to everybody, implies that the biophors have only 
secondarily acquired historic qualities, and that they are 
primarily concerned in the production of the fundamental 
processes of life. We may therefore dismiss for the present 
purpose the complications introduced by heredity and con- 
fine our inquiry to the functions of biophors as bearers of 
the essential vital qualities. 

It is urged in favour of a theory of biophors that life 
must be connected with a material unit of some sort (Weis- 
mann) ; that function presupposes structure (Whitman), and 
that the structure necessary for the exhibition of such 
complicated functions as those of living protoplasm cannot 
be of such a simple molecular kind even as the micellar 
structure postulated by von Sachs and Niigeli, but must 
consist of a definite arrangement, an architecture or organ- 


isation of separate living particles, the aggregate functions 
of which produce the vital phenomena. It is further urged 
in favour of this view of organisation, that in almost all 
cells we are able to recognise structures under the micro- 
scope each of which behaves in respect of growth and 
multiplication in a manner analogous to that in which the 
cell behaves. Not only the nucleus but also the chromatin 
bodies, the microsomata of which these are composed, the 
centrosomes, the green chromatophores of plant cells, may- 
be observed to increase in size, i.e., to grow and to multiply 
by division, and it is held that this is proof that the ultimate 
particles composing these bodies must assimilate, grow and 
divide in a manner similar to that in which cells assimilate, 
otow and divide. 

This view, whilst receiving a considerable measure of 
support from other sources, has been most energetically 
supported by Wiesner, 1 whose extensive work on the subject 
has received the weighty approval of Weismann. Wiesner 
refers in detail to the various structures in the form of 
granules or corpuscles which may be observed in animal 
and vegetable protoplasm, and he attributes to one and all of 
them the powersof assimilation and multiplication bydivision. 
Nor does he confine himself to the living substance gener- 
ally recognised under the name of protoplasm. He labours 
at great length to prove that the cell wall, so often con- 
sidered as an inert non-living product of living protoplasm, 
is not in fact dead, but contains a living substance distin- 
guishable under the name dermatoplasm, and ultimately 
composed of structural elements of the same fundamental 
nature as that of the cytoplasm. These ultimate particles 
are the fi/asomes, which form the central point of his theory 
of the constitution of living matter. Further than this he 
accepts in full the theory of De Vries with regard to vacuoles, 
and considers them to be just as much independent organ- 
isms as the chromosomes, the centrosomes, the chlorophyll 
bodies and other things. This theory of vacuoles, which 

1 J. Wiesner, Die Elementar structur und das Wachsthum der Lebendem 
Substanz. Wien : Alfred Holder, 1892. 


assumes that they are products of minute bodies called 
tonoplasts, is of itself improbable, and is contrary to the 
teaching of observations which may readily be made on the 
constitution and behaviour of vacuoles in living protoplasm. 
It has been shown by Butschli l that the contractile and 
other vacuoles of Protozoa continually make their appear- 
ance without owing their origin to the division of previously 
existing vacuoles. It is not possible to go into details here, 
but the reader will find a full discussion of this question in 
Butschli's work (p. 230) as also a resume of the various 
theories which have from time to time been put forward on 
the subject of the granular theory of protoplasm. But even 
if peculiar views on the nature of vacuoles be laid aside, 
the gist of Wiesner's arguments is not materially weakened. 
None of the structures which are observable in protoplasm 
are observed to originate neogenetically : they are all, he 
says, derived directly by division from pre-existing struc- 
tures of similar character. In short, he fully accepts the 
aphorism put forward somewhat earlier by Altmann : 
" Omne granulum e granulo ". Wiesner does not definitely 
say that the various particles observable in protoplasm are 
to be severally identified with the ultimate vital units, his 
plasomes. Some of them may be individual plasomes, but 
the majority of them are, he thinks, aggregates of plasomes, 
units of a hioher order which in turn are combined to form 
the still higher unit the cell. Thus he presents a scheme 
of organisation which, instead of taking the cell as the 
lowest structural unit, goes several grades lower ; instead 
of the old conception of 

organ — tissue — cell, 

he represents the scheme of organisation as being 

organ — tissue — cell — granules — plasomes. 

A detailed criticism of Wiesner's views would occupy a 

much larger space than I have at my disposal, and such a 

criticism is unnecessary, since all that need be said has 

1 O. Butschli, Investigations on Microscopic Foams and Protoplasm. 
Authorised translation by E. A. Minchin. London : A. and C. Black, 


already been set forth by other authors in their criticisms of 
similar theories, particularly by Biitschli {loc. cit., p. 195) 
and O. Hertwig, both of whom occupy themselves with 
Altmann's views, which are to all intents and purposes 
identical. Only a few of the most important points need be 
touched upon here. 

It is certainly a remarkable fact, and confirmed by 
abundant experience, that many of the constituent parts of 
cells are produced by divisions which recall the divisions of 
the cell itself. The nucleus is the most important and the 
most familiar constituent of the cell : it is within the experi- 
ence of every biologist that nuclei are never observed to 
originate neogenetically, but always by division of a pre- 
existing nucleus. The chromatin elements of the nucleus may 
be shown to be composed of minute particles, the so-called 
chromosomes, and these reproduce themselves by division, 
and are never observed to originate neogenetically. The 
same statement holds good for the centrosomes, for 
chlorophyll corpuscles and for various kinds of chroma- 
tophores. It is not to be denied that these facts, which 
become more and more familiar to the working microscopist, 
appear to lend a powerful support to the theory of 
biophors ; in a limited sense they may be said to be a proof 
of the statement that the cell is an organised body. 
Whether, as Wiesner claims to be the case, there are many 
other constituents of cells which similarly reproduce their 
kind by division, and are never observed to originate 
independently, may for the present be left out of considera- 
tion. The evidence that amylum grains and granules of 
various kinds behave like the centrosomes in this respect, is 
as yet too slight, and the observations are too conflicting to 
enable us to come to a judgment without entering into a 
mass of detail which is not wholly relevant to the question 
at issue. But there is at least one criticism which is worthy 
of mention, namely, that of Biitschli, who points out {loc. cit., 
p. 200) that among the strongly staining granules of proto- 
plasm there are bodies which are not actually constituents 
of the protoplasm but are symbiotic micro-organisms. 
The existence of such organisms, which have been called 


Bacterolds, has been frequently demonstrated in animal 
and vegetable cells, and Biitschli points out that granules 
similar in appearance to bacteroids occur in the Vorticellinse 
and may be observed at certain times to be in a state of 
rapid proliferation. 

Just before writing these pages I have been shown 
preparations exhibiting the numerous bacteria symbiotic 
in Pelomyxa, and it is very possible that the rapid 
proliferation of bacteroids has been mistaken by various 
observers for the reproductive activity of granules forming 
an integral part of the cell. It is almost certain that the 
mistake has been made in some cases, and until further 
investigation has increased our knowledge of the various 
micro-organisms which are symbiotic or parasitic in cells, it 
is well to be somewhat sceptical of statements regarding 
the divisional processes of cell contents. It would seem then 
that our present knowledge does not justify our regarding 
all the particles of a cell as originating in a similar manner 
from the division of pre-existent similar particles, though 
we must affirm in the most positive manner that some few 
of the constituents of the cell originate in this way only, and 
are never produced de novo. The question now to be 
considered is this : must we, because these bodies (the 
centrosomes, chromosomes, etc.) assimilate, grow and 
reproduce themselves by division, regard them as indepen- 
dent vital units? A cell exhibits these phenomena and the 
cell is regarded as an independent unit in posse, if not 
actually in esse ; must we therefore attribute to all bodies 
exhibiting the same phenomena the character of indepen- 
dent units ? The answer, I believe, is very decidedly no. 
Cells would never have been regarded as independent units 
if they had merely been observed to assimilate, grow and 
divide, whilst retaining their connection with other cells 
undergoing the same processes. The quotations which 
I have given from Schwann's work show that the 
theory that the cell is an independent life unit was not 
founded on the fact that it assimilates, grows and divides, — 
Schwann indeed overlooked the phenomena of reproduction 
by division — but upon the fact that cells are capable of 


leading an independent existence. This is so important a 
part of the cell-theory that I may again quote in his own 
words Schwann's reasons for calling the cell an elementary 
unit of life. " Now if we find that some of these elementary 
parts not differing from the others are capable of separating 
themselves from the organism and pursuing an independent 
growth, we may thence conclude that each of the other 
elementary parts is already possessed of the power to take 
up fresh molecules and grow, and that therefore each 
elementary part possesses a power of its own, an indepen- 
dent life." 

In the case which we are considering the very faculty 
which was so powerful a reason for regarding cells as 
independent units is wanting. Nobody has ever observed 
a nucleus or a centrosome or even a chromatophore to 
separate itself from the cell and pursue an independent 
existence. And not only is there no recorded case of the 
constituent particles of cells separating themselves spon- 
taneously from the cell, but experiments which have been 
made with the express purpose of determining whether 
these particles can live apart from the cell to which they 
belong have in every case given a negative result. Even 
the nucleus, highly complicated as it is, and itself composed 
of smaller particles which may easily be demonstrated, 
perishes when removed from the cell body. The chroma- 
tophores similarly perish, and so no doubt would the 
centrosomes if it were possible to isolate such very minute 
particles. Many instances might be cited in proof of this, 
but it is scarcely necessary to bring forward the details ; 
the reader can obtain them by reference to the works of 
Nussbaum, 1 A. Gruber and Verworn. 2 

It is of some interest in this connection to contrast the 
process of reproduction in unicellular and multicellular or- 
ganisms. In the latter reproduction is effected by the 
separation of a single unit, a cell, from the aggregate, and 

1 M. Nussbaum, Biol. Centralblatt, vol. iv. 

2 Max Verworn, " Die physiologische Bedeutung des Zellkerns," 
Pfliiger s Archiv, vol. li., 1892. 


the unit so separated has from the time of its separation an 
independent individuality and eventually reproduces the 
aggregate. The fact that the union of two cells is commonly 
necessary for the maintenance of life and the exhibition of 
the powers of development, need not be urged as an objec- 
tion to this simple statement of the case, for the facts of 
parthenogenesis show that the union of two cells is not an 
essential feature. Now if we adopt Wiesner's scheme, and 
imagine that organisation does not stop at the cell, but that 
beyond this there are granules, and beyond these again 
plasomes, and that the plasomes stand in the same relation 
to the cell that the cell stands to the multicellular organism ; 
we should expect to find that in the reproduction of 
monocytial organisms the plasome plays a part anal- 
ogous to that played by the cell in the reproductive 
processes of polycytial organisms. But we find nothing 
of the kind. The monocytial organism reproduces itself in 
just the same way as the polycytial, by the separation of a 
cell, complete in all its parts. There is no such thing 
known, even in cases where a flagellate or a radiolarian 
breaks up into innumerable particles or spores of extreme 
minuteness, as the separation of any one individual con- 
stituent of a cell possessed of the power of leading an in- 
dependent existence and in time of reproducing all the other 
constituents. Every spore, however minute, has its portion of 
the cytoplasm and its share of nuclear matter. If there are 
any other constituents, it probably has its share of these 
also, but one cannot speak with certainty on this point, for 
positive evidence is wofully deficient. At any rate Wiesner, 
holding fast to his theory that nothing, not even an amylum 
or an aleurone grain, is produced neogenetically, is at great 
pains to prove that in cellular reproduction all the parts 
of the parent are transferred to the offspring. Assuming 
that this is so, and remembering that there is abundant 
evidence that nuclear matter and cytoplasm are always 
transferred, it is evident that the relation in which the 
plasomes or biophors, regarded as ultimate vital units, 
stand to the cell, is not at all the same as the relation in 
which the cell, regarded as an ultimate unit, stands to the 


polycytial organism. Biitschli, in a short but very weighty 
sentence, 1 brings forward the same argument that I have 
just used in opposition to Altmann's theory of the part 
played by granules in the vital processes of protoplasm. In 
my judgment the argument as far as it goes is a sound 
one, but I am aware that it does not altogether refute the 
theory of biophors, but only that part of it which states that 
as cells are to polycytial aggregates so are biophors to cells. 
This refutation, however, seems to me to be a considerable 
gain. For it enables us to apprehend that the structure or 
constitution of the cell, whatever it may be, is not to be ex- 
pressed in the same terms as the structure of the higher 

It may be objected that nobody does express the 
structure of the cell in such terms, but the objection does 
not hold good. It is true that most authors are more 
guarded in their expressions than Wiesner, and evade the 
responsibility of declaring that the biophor is to the cell 
as the cell is to the polycytial organism, by means of re- 
servations, couched for the most part in terms so ambiguous 
and even transcendental that the whole issue is involved in 
an obscurity from which it seems hopeless to try to escape. 
But these expedients are really of little use. The fact re- 
mains that in every case the fundamental idea is the same, 
that the phenomena exhibited by isolated cells having an 
independent individual existence are of essentially the same 
kind as the phenomena exhibited bypolycytial organisms and 
must be explained on the same grounds. 

If it be not so, what is the meaning of the argument 
which was first put forward in definite shape by Brlicke, 
and has been repeated by every author who attacks the 
question in the same manner that he did, that the com- 

1 " So long as the individual constituents of the cell are not seen to 
persist when isolated, nor are distinct living phenomena observed in them, 
it is very dangerous to speak of their life as something which they possess 
in themselves. They are so far living, as long as the opposite is not proved, 
in that they are parts of living organism, so that the granula may be 
living in the same way as the nucleus, even though they no longer betray 
any sign of life after isolation" (Joe. eit., p. 199). 



plexity of the phenomena exhibited by individual cells, say 
by an amceba, is so great, the functions observed are so 
many and so various in their kind that they can only be 
explained by the assumption that protoplasm is an organised 
body ? Taking the words of O. Hertwig as a fair ex- 
pression of current opinions on the life processes of a poly- 
cytial organism, " that the aggregate life processes of a com- 
posite organism appear to be nothing more than the ex- 
ceedingly complicated result of the individual life processes 
of its numerous and variously functional cells," it is evident 
that to the minds of Briicke and his successors the aggre- 
gate life processes of the corpuscle of protoplasm called a 
cell are nothing more than the highly complicated result of 
the individual life processes of its numerous and variously 
functional biophors. If they do not mean this, I am quite 
at a loss to know what they do mean, or to understand the 
relevancy of the so-called axiom laid down by Whitman, 
that "function presupposes structure," or the meaning of 
the statement expressed so often and with such obvious 
satisfaction, that " the cell is an organism ". These sen- 
tences, so terse and so epigrammatic, exercise a peculiar 
fascination over most minds. To understand their exact 
applicability to the question at issue they must be carefully 
examined. Function presupposes structure. To the bio- 
logist who makes a rapid mental survey of his experiences, 
this appears to be a generalisation of universal truth. 
Physiology, which draws its inferences almost exclusively 
from the study of the higher animals, tells us that ultimately 
every function of the composite organism is to be referred 
to a particular group of cells, and that cells differ in kind 
according to the different functions which they exhibit. So 
much is this truth forced upon us that if conceivably a new 
function were to make its appearance, we should immediately 
search for the cell groups appropriate to the performance of 
that function. So far so good, but before proceeding further 
we must take note that the statement that function pre- 
supposes structure is a generalisation founded on experience. 
It is not an axiom as Whitman calls it, for an axiom is a 
proposition which is self-evident, and this assuredly is not. 


The next step is to transfer this generalisation, founded on 
experience, into a new region, to the functions of cells. In 
order to do this we should possess the same experiences 
with regard to the functions of cells which we possess 
with regard to the functions of composite organisms. But 
these experiences are entirely wanting. We observe that 
protoplasm exhibits functions, that it assimilates, that it is 
irritable, that it is contractile, that it is reproductive, and so 
forth ; but who has been able to demonstrate or even to 
suggest with any plausibility that there are structures 
specially devoted to assimilation, to contractility, to irrit- 
ability, and to reproduction in protoplasm? It is evident 
that the absence of any such experiences has been felt by 
many observers, who have accordingly studied protoplasm 
with a view to finding the required structures, and some 
are inclined to say that the nucleus or perhaps the centro- 
some is reproductive, the amylum and aleurone bodies are 
assimilative and so forth. But there are protozoa endowed 
with active functions which have no centralised nucleus ; 
the presence of centrosomes has yet to be demonstrated in 
protozoa in general, and there are forms in which, as 
Biitschli well points out, the protoplasm is homogeneous, 
e.g., in the pseudopodia Gromia dujardini} The reader 
should refer to Biitschli's work for a discussion of the sub- 
ject of hyaline protoplasm (loc. cit., p. 262). The fact that 
it exists is of the highest importance, for it shows that there 
is living substance exhibiting the usual vital phenomena of 
assimilation, contractility, etc., which, nevertheless, defies 
all attempts to recognise an organisation which in the light 
of previous experience would seem adequate to the effects 
produced, and it shows also that the centrosomes, the 
amylum grains, and their analogues, and the whole category 
of granules are secondary phenomena, which may be 

1 Not only are there no granules in homogeneous protoplasm, but the 
alveolar structure of it is unrecognisable. It is easily shown, however, that 
the homogeneous substance is produced from alveolar protoplasm and is 
capable of reconversion into it. The physical explanation of the dis- 
appearance of the alveolar structure is given by Biitschli on p. 264 of the 
English translation of his work. 


altogether absent and yet the life processes go on un- 

It must be confessed then, that the experiences which 
so amply justified our generalisation when applied to com- 
posite organisms are altogether lacking when we seek for 
a justification for applying it to the simplest unicellular 
organisms. Moreover I have just shown that in one im- 
portant particular at least, we do not merely lack these 
experiences, but that we have experiences of an entirely 
different kind. In face of this is it not obvious that the 
captivating generalisation must be abandoned altogether in 
the region which we are now discussing ? For it is founded 
on experience, and where experience fails or is contradictory 
the generalisation fails also. 

After what has already been said it is unnecessary for 
me to enter into a detailed examination of the other state- 
ment which is considered to mark a great advance in bio- 
logical thought, that "the cell is an organism". It is 
sufficient to say that if this proposition means anything at 
all, it means that the cell has an organisation which is 
similar in kind to that of a composite organism of which a 
cell is a part. If I am told that it does not mean this, but 
something else, then I ask, firstly, what does it mean ? And, 
secondly, if it does not mean this, what necessity is there for 
assuming that the protoplasm of the cell is built up of bio- 
phors, the biophor being the elementary living constituent, 
assimilating, growing and dividing, taking up definite posi- 
tions in the cell, combining with others like or unlike itself to 
form higher aggregates, and so impressing a fixed archi- 
tecture on the cell of which it is a component ? Why, in 
short, if the statement does not mean that the organisation 
of the cell is the same in kind as the organisation of a 
composite animal, why then does everybody who believes 
that the cell is an organism ascribe to it an organisation 
which is the same in kind as that of the higher animals ? 

The fact is, and it is patent to everybody, that most 
authors do conceive of the cell-organisation as being the 
same in kind as the organisation of higher animals. They 
either have the courage of their opinions, like Wiesner, 


and say so in so many words, or they tacitly admit it by 
their description of what they conceive cell-organisation to 
be. They are dominated by the cell-theory. Mr. Adam 
Sedgwick has recently said that the cell-theory is an incubus 
which perverts the minds of biologists, whose minds are so 
saturated with conceptions borrowed from the cell-theory 
that they are unable to see anything else. I have else- 
where found fault with this statement, but when the theories 
of cell-organisation are considered, I must freely confess 
that he has right on his side. Not only does the zoologist 
believe "that the cell is the unit of structure, and that it 
forms the basis of organisation in the metazoa," but he also 
believes that some correlative of the cell forms the basis 
of all organisation whatsoever. His eyes are " blinded to 
the most patent facts" by ideas derived from the cell-theory, 
and it is not too much to say that the theory does " obstruct 
the way of real progress in the knowledge of structure ". 
Whether consciously or unconsciously the believer in bio- 
phors starts with ideas derived from the cell-theory, he 
tacitly assumes the universal applicability of the proposition 
that function presupposes structure, and he seeks to explain 
the functions of protoplasm by attributing to it an organisa- 
tion which in all essential characters is the equivalent of 
the organisation of the metazoa. Since I have just shown 
that there is no justification for transferring a generalisation 
based upon experience to a region in which experience is 
either wholly wanting or, if present, of a different kind, it is 
hardly necessary for me to elaborate and show that it is 
equally unjustifiable to attribute to the unknown a plan of 
organisation identical in kind with the plan which we have 
learnt by experience to recognise as the attribute of the 

Some time ago I pointed out that there was a fallacy in 
the word organism. 1 Whitman has ridiculed the statement, 
yet the more I reflect upon it the more I am convinced 
that the fallacy exists, and that it is in the highest degree 
mischievous and misleading. By an organism we mean 

1 G. C. Bourne, " Epigenesis and Evolution," this journal, vol i., 1894. 


either an independent living thing, in which case the term 
is loose but applicable to every animal in the monocytial 
stage, or we mean a thing possessing organisation, and by 
organisation we mean a certain structural plan, the idea of 
which is a generalisation from our experience of animal and 
vegetable structure in general. That this is historically 
and in fact the connotation of the term organisation is in- 
dubitable. 1 When we use the term organisation we 
either use it in this connotation or in some other. If we 
use it in the same connotation with respect to protoplasmic 
structure, we are consistent, but, as I have shown, we are 
applying ideas derived from one set of phenomena to 
another set of phenomena to which they are not appropriate. 
But if we use it with another connotation, then we expose 
ourselves at once to the risk of the well-known fallacy which 
is inseparable from the use of the same term with different 
connotations. If the two connotations are clearly defined 
and generally understood, the fallacy may be avoided, 
though the inconvenience remains ; but if the one connota- 
tion is clear and definite whilst the other is vague and 
ambiguous in the highest degree, no amount of circum- 
spection will prevent our falling into the fallacy almost at 
the first opportunity. This is exactly the case with the term 
organisation. In the one sense we know its connotation 
exactly, and when authors use it in that sense they have, 
in the course of their arguments, to adhere strictly to the 
technical sense of the word. Most of them do this, for 
they are aware of the absurdities and inconsistencies into 
which they would fall if they did otherwise. But what 
of those who use the term with another connotation ? They 
assure us that it does not denote a plan of structure like in 
kind to that of the metazoa : what then does it denote ? 
Something so vague, so unreal and unsubstantial that we 

1 Thus in Worcester's Dictionary of the English Language, 1881 : — 
Organisation. The condition of an organised body or the totality of 

parts which constitute and the laws which regulate an organised body. 

Organised. Formed with organs : composed of several individual parts 

or organs, each of which has its proper function and conduces to the 

existence of the entire system. 


are even at a loss to know to what to apply it ; its connota- 
tion has never even been attempted. The futility of using a 
term without connotation and with the most vague denotation 
is so well illustrated by the following passage from Whit- 
man that I cannot refrain from introducing it here : " When 
we speak of the organisation of the germ as cut directly 
from a pre-existing parental organisation of the same kind 
we are not thinking of the definitive organisation which 
belongs to the fully formed organism, but of that primary 
organisation which belongs to the protoplasm itself". This 
raises our expectations, we are going to hear something of 
the primary organisation which belongs to protoplasm itself. 
Whitman continues: "We are so accustomed to connect 
the idea of organisation with the anatomical organs of the 
adult that we are apt to forget that there is a primary 
organisation which underlies every anatomical organ. The 
germ has this primary organisation ; it is therefore an 
organism, and as such may dominate its own development." 
From which weighty and sententious passage we gather 
that the germ is an organism because it has a primary 
organisation which is not the definitive organisation which 
belongs to the fully formed organism, but a primary organ- 
isation which belongs to protoplasm itself. What on earth, 
we may well ask, is this primary organisation ? The 
answer is given on the same page. It is "that original 
constitution of the germ which pre-determines its type of 
development and the form which ultimately distinguishes it 
from other species developing under like external conditions". 
The terms "original constitution" and "primary organisa- 
tion " are merely synonyms. So we learn that the primary 
organisation so important to those who have more thought- 
fully scanned the gap between the cell and the physical 
molecule, is the primary organisation of the germ, which 
pre-determines its type of development, etc. I hope that 
others are satisfied by this most remarkable piece of 
scientific exposition. For myself I must humbly confess 
that I am none the wiser for it, any more than I should be 
if I asked what was a Megalosaurus and I was told : "A 
Megalosaurus, why you know it is a big lizard, it is — a — a 


Megalosaurus in fact ". Nor is confusion less when I am 
told in one sentence that the organisation of the germ cut 
directly from pre-existing parental organisation of the same 
kind is not the definitive organisation which belongs to the 
fully formed organism, but is that primary organisation 
which belongs to protoplasm itself, and I read in the 
sentence immediately preceding that " the essential thing is 
not simply continuity of germ substance of the same 
chemico-physical constitution, but actual identity of germ- 
organisation with stirp-organisation ". The organisation of 
the germ is identical with the organisation of the stirp, and 
yet the organisation of the germ is not that of the fully 
formed organism, but is a primary organisation which 
belongs to protoplasm itself. What does it all mean ? It is 
different and yet it is identical, and it is organisation, organisa- 
tion, toujours organisation. I beg Dr. Whitman, for pity's 
sake, to descend from his altitude, scarcely dreamed of in the 
philosophy of Harvey and Wolff, and to condescend to 
inform a poor bewildered mortal, who confesses to a pre- 
judice in favour of things which he can understand, what 
this wonderful primary organisation is. 

Seriously speaking I believe that organisation either 
means a plan of structure of the same type as the structure 
of higher animals and plants, and capable of being described 
in intelligible terms as it has been by Weismann, Wiesner 
and others, or it means nothing at all ; it is a mere phrase 
which seeks to cover but does not conceal our ignorance. 

G. C. Bourne. 
( To be continued. ) 


IF we define solutions as homogeneous mixtures of sub- 
stances in variable proportions, we are at once obliged 
to admit the existence of solid solutions, for there are many 
mixed solids which fulfil the requirements of this definition. 
Common potash alum, for example, can crystallise together 
with ammonia alum, and form mixed crystals which are 
perfectly homogeneous and of the same composition through 
out, although the proportions of the two constituents may 
be varied at will by proper selection of the aqueous solution 
from which the crystals separate. 

We are inclined, however, to look in solutions for some- 
thing more than mere homogeneity and uniformity of com- 
position, and perhaps one of the most obvious characters of 
a liquid solution is this, that should it at first be of different 
composition in different parts of its mass, there is always 
present the tendency of the dissolved body to attain a uni- 
form distribution throughout the solvent. The process of 
equalisation of the composition, or diffusion, occurs in all 
solutions which are more concentrated in one part than in 
another, the dissolved substance moving from the place of 
greater to the place of less concentration. Diffusion in 
solution goes forward very slowly if the liquid is protected 
from mechanical disturbance and sudden change of tem- 
perature, months being requisite for the attainment of uni- 
form concentration if a comparatively short column of pure 
solvent is placed above a denser layer of strong solution 
contained in the bottom of a cylinder. If diffusion takes 
place in solids we might expect it to proceed even more 

A class of substances which form in some sort a connect- 
ing link between liquids and solids, and are specially suited 
to the study of diffusion phenomena, is to be found in jellies. 
Graham, to whom we owe our first exact knowledge of 
diffusion in liquids, prepared a stiff jelly containing common 
salt in solution in one part, and compared the rate at which 


the salt diffused in it with the rate at which salt dif- 
fused in pure water. He found that the diffusion in the 
jelly took place almost, if not quite, as fast as in water 
itself. The composition of the jelly was 2 per cent, gelose 
and 98 per cent, water, so that, as far as actual substance 
was concerned, the salt had to meet practically the resistance 
of water alone in both cases, and the experiment showed 
that the mere change in apparent condition of the whole 
mass had little or no influence on the rate of diffusion. 
Subsequent experiments have served to confirm Graham's 

When we pass to solids proper we find that instances 
are not wanting of what is apparently diffusion within them. 
Van't Hoff in his fundamental paper on solid solutions gives 
numerous examples. In the preparation of steel by the cem- 
entation process bars of wrought iron are packed in charcoal 
and subjected to a red heat for several days. The charcoal 
gradually penetrates the iron and converts it into steel. It 
matters little for our purpose what the particular form is that 
the carbon assumes during its passage through the iron — in 
some fashion or other it reaches the centre of the dense bar. 
The distribution of the carbon, too, if the operation is inter- 
rupted before uniformity has been attained, is precisely what 
would be expected if the phenomenon were one of real diffu- 
sion ; and the influence of time is the same in both processes. 
Not only has carbon been observed to pass through iron, 
but it has even been proved to travel slowly through por- 
celain, when porcelain crucibles have been heated in a bed of 

When a metal such as copper is deposited galvanically 
on another metal, it penetrates beyond the surface of the 
latter into its substance, and zinc objects which have been 
lightly coppered are, even when protected by a coating of 
varnish, occasionally observed to become white again 
owing to the gradual mixing of the two metals near the 

Professor Spring, of Liege, who has devoted special 
attention to the chemical behaviour of solids under high 
pressure, has supplied some interesting instances of pheno- 


mena which can only be explained by the assumption of 
solid solutions. When equivalent proportions of barium 
sulphate and sodium carbonate are finely powdered, 
intimately mixed, and subjected to a very high pressure, a 
double decomposition takes place with formation of barium 
carbonate and sodium sulphate. The decomposition, how- 
ever, is not complete, only 20 per cent, of the original 
substances being transformed. If, on the other hand, we 
start with a mixture of barium carbonate and sodium 
sulphate and compress it, we find that the reverse trans- 
formation now occurs, barium sulphate and sodium carbonate 
being formed, and that to the extent of So per cent, of the 
original substances present. Here we are evidently dealing 
with a state of equilibrium between the four substances 
above mentioned, which can only exist together permanently 
under pressure in certain definite proportions. If these 
proportions are departed from, the system so transforms 
itself that the requisite state for equilibrium is attained. 
Now this of itself points to the substances existing here in 
a state analogous to that of bodies in liquid solution, for we 
know that in general definite proportions are necessary in 
solutions for stable equilibrium to exist. In the case of 
solids the general rule is that when they are in equilibrium 
under given conditions in one proportion, they are in 
equilibrium under the same conditions in every other pro- 
portion. The behaviour, then, of these solids under pressure 
is analogous to the behaviour of substances in solution, and 
different from the ordinary behaviour of solids. The con- 
tinuance of the pressure is not essential to the establishment 
of such a definite solid equilibrium, for Spring has shown 
that by relieving the pressure after 73 per cent, of a 
system of barium carbonate and sodium sulphate had been 
transformed, the process continued, though less rapidly, 
and after a week had reached the proportion of 80 per cent, 
necessary for equilibrium. Here diffusion must have played 
a part, for no matter how finely divided the reacting sub- 
stances originally were, their surface of contact (where alone 
the mutual decomposition could take place if there were no 
diffusion) must have been comparatively small. 


It is well known that some metals have the property of 
allowing certain gases to pass through them under favourable 
conditions, the most thoroughly investigated instance of 
this kind being the permeability of the metal palladium to 
gaseous hydrogen at moderately high temperatures. At 
about 300° C. hydrogen can pass quite freely through a 
palladium septum, and it is difficult to conceive the nature 
of this phenomenon without admitting the existence of 
diffusion in the solid. Whether the hydrogen is dissolved 
in the palladium or forms a compound with it, as has been 
asserted, is of little consequence, for in the latter case the 
compound superficially produced must have possessed the 
power to penetrate the remaining metal, or to allow of the 
passage of hydrogen through itself. 

Connected with the process of diffusion in solution we 
have the phenomena of the conduction of electricity in 
solutions, or electrolysis. Here the electric current is 
carried by material particles, and the resistance that these 
experience in their passage through the solution is of the 
same nature as the resistance offered to diffusion. Helm- 
holtz, in his Faraday lecture, drew attention to the fact that 
glass behaves as an electrolyte towards an electric current, 
i.e., that the current in passing through the glass is as- 
sociated with two currents of particles moving in opposite 
directions. The particles travelling towards the negative 
pole of the battery have since been proved to move faster 
than those moving towards the positive pole. Lehmann 
also has shown that when two silver electrodes are immersed 
in fused iodide of silver, which is afterwards allowed to 
solidify, and a current of electricity is passed through the 
solid iodide, one of the electrodes increases in weight at the 
expense of the other, and that the phenomenon can be 
reversed by reversing the current. 

These examples will suffice to indicate that we are not 
without data to establish an analogy between the behaviour 
of certain solids and the behaviour of ordinary liquid solu- 
tions. Since the appearance of van't Hoff's original paper 
on the subject a considerable number of researches have 
been published more or less directly bearing on the question, 


but the results achieved have on the whole been small, owing 
chiefly to the experimental difficulties encountered. 

An important application of the idea of solid solutions 
was made by van't Hoff in explaining the abnormalities 
that are sometimes met with in the determination of mole- 
cular weights by the lowering of the freezing-point in solu- 
tions. It had been proved theoretically that the freezing- 
point of a given solvent should be depressed to a certain 
value (calculable from the freezing-point and the latent 
heat of fusion of the solvent) when the solution was of 
normal concentration, i.e., contained one gram-molecule 
of dissolved substance per litre. The nature of the dis- 
solved substance should be without influence on this value. 
Now, whilst it was ascertained experimentally that this 
theoretical relation was in the vast majority of cases ac- 
curately fulfilled, yet there remained certain combinations 
of dissolved substance and solvent which gave values of the 
depression constant altogether at variance with the cal- 
culated value. Thus, metacresol dissolved in phenol gave 
a depression of 48 instead of 74, and thiophene dissolved in 
benzene a depression of 34 instead of 53. Van't Hoff's 
explanation of these and similar abnormally low values of 
the depression was that the freezing-point observed was not 
in the strict sense the freezing-point which had been assumed 
in the theoretical reasoning. The true freezing-point of a 
solution is the temperature at which the liquid is in equi- 
librium with the solid solvent. The freezing-point of an 
aqueous salt solution, for example, is the temperature at 
which it can exist in contact with pure ice without the ice 
melting or without fresh ice being deposited from the solu- 
tion. Now, in the exceptional cases above alluded to it is 
known that the solid and the solvent have a tendency to 
crystallise together, i.e., to form mixed crystals, so that the 
substance that separates out is not the pure solvent but 
rather a solid solution. The temperature at which such a 
solid solution would be in equilibrium with the liquid solu- 
tion might not by any means be the freezing-point of the 
solution as above defined. The apparent observed freezing- 
point of the solution, therefore, would not in general coincide 


with the calculated depression, and van't Hoff from theo- 
retical considerations showed how the divergence could be 
estimated from a knowledge of the composition of the solid 
which actually separated out from the solution on cooling. 
That the abnormal values for the points of solidification 
depend on the separation of the dissolved substance along 
with the solvent has now been experimentally verified in a 
considerable number of cases. Heycock and Neville found 
that for the case of solutions of antimony in molten tin, the 
freezing-point of the tin was raised instead of lowered by 
the presence of the second metal. Kiister has shown that 
this and similar instances are susceptible of a very simple 
explanation. The two metals separate out together in very 
nearly the same proportion as that in which they remain 
behind in the liquid, so that the solution solidifies as a 
whole. In such circumstances the point of solidification of 
the liquid can be calculated by the simple mixing formula. 
If the melting-point of each pure substance is multiplied by 
the proportion in which it exists in the mixture, the sum of 
the two numbers thus obtained will give the point of solidi- 
fication of the solution. As antimony melts 200 degrees 
higher than tin, the admixture of the former in however 
small proportion will, since the mixture freezes as a whole, 
raise the point of solidification instead of lowering it, as 
would be the case if pure solid tin separated from the liquid 
on cooling. 

Not only do solutions exhibit a lower freezing-point than 
that of the pure solvent, but they also exhibit a lower 
vapour tension. The pressure of aqueous vapour over 
salts containing water of crystallisation may in many cases 
be measured with accuracy, and there it is found that the 
isomorphous admixture of another salt lowers the vapour 
pressure of water which is in equilibrium with the solid. 
Thus the vapour tension of a mixed crystal of ordinary alum 
with iron alum is less than the vapour tension of either of 
its components. In this respect then the mixed crystal 
behaves as a solid solution. Again, the solubility of a sub- 
stance is diminished when it itself acts as a solvent for 
another substance insoluble in the original solvent. Of the 


three liquids, ether, water, and benzene, ether and water are 
partially miscible, benzene and water are immiscible, and 
ether and benzene miscible in all proportions. Suppose we 
take water as the original solvent — then on shaking it up 
with ether we find that the latter dissolves to a certain 
definite extent in it, i.e., possesses a certain solubility 
in water. If now we previously dissolve benzene in the 
ether which we shake up with the water, we find that the 
water will now take up less ether than before. The solu- 
bility of ether in water is thus diminished when benzene is 
dissolved in it — and this behaviour is characteristic of all 
such combinations of substances. 

A case of this kind where two solids play the part of the 
ether and benzene in the previous instance has been 
thoroughly studied by F. W. Klister. The solid hydrocarbon 
naphthalene is, like the hydrocarbon benzene, insoluble in 
water ; /3-naphthol, on the other hand, is, like ether, 
sufficiently soluble in water to permit of accurate estimation. 
But naphthalene and /3-naphthol can crystallise together in 
any proportion so as to form a complete series of isomor- 
phous mixtures, the melting-points of which vary according 
to the rule given above for mixtures of antimony and tin. 
A comparison of the amount of /3-naphthol dissolved by a 
given quantity of water from such mixtures led to some- 
what unexpected results. Instead of the addition of a 
small quantity of naphthalene to /3-naphthol lowering the 
solubility of the latter in water, it was found that mixtures 
containing as much as 30 per cent, of naphthalene had 
precisely the same solubility as /3-naphthol itself. As more 
naphthalene was added the solubility increased slightly, 
afterwards to diminish continuously to zero as the mixture 
was made to contain more and more naphthalene. The 
explanation of this behaviour suggested by Klister is that 
naphthalene and /3-naphthol are capable of forming a chemi- 
cal compound consisting of one molecule of each substance, 
this compound being decomposable by water, an assumption 
by no means improbable, as many similar cases have been 
observed. If we allow further that the solubility of the 
compound is greater than the solubility of /3-naphthol, the 


results are satisfactorily accounted for. The diminution 
of solubility when much naphthalene is present is the 
normal depression of the solubility of the compound by 
the addition of excess of naphthalene. The solubility 
greater than that of pure naphthol is the solubility of the 
compound naphtholnaphthalene. The constant solubility 
(equal to that of /3-naphthol) observed when there is little 
naphthalene in the mixture is the solubility of /3-naphthol, 
for the naphthalene in the mixture is in the form of the 
compound naphtholnaphthalene, which is decomposed at 
the surface by water into naphthalene and /3-naphthol, 
which exist now alongside of each other and not in the 
intimate union of a crystalline isomorphous mixture. 

In connection with the results of these experiments 
Klister is inclined to make a distinction between crystalline 
isomorphous mixtures and solid solutions proper, because in 
the former there is practically no diffusion owing to what 
may be termed the rigidity of the crystalline structure. 
He admits, however, that no absolutely sharp line can be 
drawn, as there are various intermediate degrees in which 
diffusion may take place. A reference to the examples of 
diffusion in solids previously cited in this paper will show 
that they all occur in amorphous bodies without any regular 

A point of considerable interest in the theory of solid 
solutions is that it affords us the possibility of determining 
molecular weights of the dissolved substances, and since in 
isomorphous mixtures we usually attribute similarity of 
molecular structure to the two components, we can also in 
this case form an estimate of the molecular weight of the 
solid solvent. From his experiments on the amount of /3- 
naphthol dissolved by water from mixtures of that substance 
with naphthalene, Kiister was able to calculate with a high 
degree of probability the molecular weight of each of these 
substances in the solid state. In the first place he found 
that with mixtures containing excess of naphthalene the 
ratio of the square root of the concentration of /3-naphthol 
in the solid mixture to the concentration in the aqueous 
solution saturated by that mixture was very nearly constant, 


varying but little with the actual composition of the mixtures 
taken. The general theory of solutions asserts that when a 
substance(here /3-naphthol) is divided between two immiscible 
solvents (here water and naphthalene, or naphtholnaph- 
thalene) it will be distributed in a constant ratio between 
the two solvents, no matter what amount of it be taken, 
provided only the molecular weight of the substance is the 
same in both solvents. In the case investigated this does 
not hold — the ratio of the concentrations in the two solvents 
is not constant ; and the molecular weight of /3- 
naphthol dissolved in water is therefore different from the 
molecular weight of /3-naphthol " dissolved " in naphthalene. 
The theory further asserts that when, as in the present 
instance, the concentration in one of the solvents is pro- 
portional to the square root of the concentration in the 
other solvent, the molecule in the second solvent must be 
twice as great as the molecule in the first. We know that 
/3-naphthol dissolved in water has the normal molecular 
weight corresponding to the formula C IO H s O ; in naphthalene 
solution it has consequently the molecular weight corre- 
sponding to the formula (C IO H 8 0) 2 . 

The theory of solutions likewise enables us to calculate 
the molecular weight of the naphthalene in the above 
experiments from the diminution of the solubility of the 
/3-naphthol in water as it dissolves more and more naph- 
thalene. In the case before us the question is slightly com- 
plicated by the existence of naphtholnaphthalene molecules, 
but Kiister was able to arrive at the result that naphthalene 
must have double the molecular weight in the state of solid 
solution that it has in the state of vapour, viz., (C IO H 8 ) 2 . 

Another well-investigated case of solid solutions is that 
offered by the absorption of hydrogen by palladium. 
T roost and Hautefeuille, in order to obtain information as to 
the state in which the hydrogen existed within the metal, 
made an extensive series of observations of the pressure of 
hydrogen in equilibrium with palladium containing different 
amounts of hydrogen. They found that with compositions 
of the solid up to one atom of hydrogen to two atoms of 
palladium the pressure of hydrogen remained constant 



at ioo° C, after which it increased rapidly as the pro- 
portion of hydrogen in the solid increased. The analogy 
between this case and the case of the solubility of mixtures 
of /3-naphthol and naphthalene in water is at once apparent. 
In both instances we have constancy of pressure (gas- 
tension) and solubility (solution-tension) within a certain 
range of composition, and then rapid variation with further 
change of composition. The conclusions arrived at in both 
instances are also similar. The constant solubility was 
attributed by Klister to the formation of a compound 
naphtholnaphthalene — the constant tension was attributed 
by Troost and Hautefeuille to the formation of a compound 
Pd 2 H, in which any excess of hydrogen was then absorbed. 
Quite recently, however, grave doubts have been thrown 
on the existence of this compound. A very careful repetition 
and extension of Troost and Hautefeuille's experiments by C. 
Hoitsema has proved that the constancy of tension observed 
by these investigators was not absolute but only approxi- 
mate, and that under slightly varying conditions the 
apparent constancy disappeared altogether. It would 
seem, therefore, that no compound of palladium and hydro- 
gen is formed when the gas is absorbed by the solid, the 
state of the hydrogen being rather one of simple solution 
in the palladium. A comparison of the concentrations of 
the hydrogen above the palladium and of the hydrogen in 
the palladium indicates that at very low pressures the 
hydrogen in the metal exists as molecules only half as great 
as those of the gas, i.e., as molecules consisting of only one 
atom. At higher pressures the concentration of the free 
gas and that in the palladium stand in a nearly constant 
ratio, from which it is to be inferred that the molecule of 
hydrogen in the metal, as well as the molecule of gaseous 
hydrogen, is represented by the formula H 2 . 

A problem which has long interested chemists is the 
determination of the nature of the process involved in dyeing. 
Some contended that the process was one of chemical union of 
the dye with the substance of the fibre, others that it was 
merely one of mechanical absorption. In 1890, however, 
O. N. Witt propounded a new theory which, on account of 


its plausibility, met with a ready acceptance in many 

According to Witt the state of the dye-stuff in the fibre 
is one of solid solution, and many analogies were advanced 
in support of this assertion. For example, dyed materials 
show the colour, not of the solid dye-stuff, but of the dye- 
stuff in solution, when there is a difference of colour between 
the two states. Solid fuchsine is green, its aqueous solutions 
are red, and so also are materials dyed with it. The dye- 
stuff rhodamine in the solid state exhibits no fluorescence, in 
solution it does, and silk dyed with rhodamine is fluorescent 
likewise. The theory of Witt thus appeared very promising 
as an explanation of the phenomena of dyeing, but a closer 
investigation has shown that it cannot be accepted uncondi- 
tionally, although some modification of it may be found to 
satisfy the experimental requirements. It has been proved 
in a considerable number of instances now investigated 
that the concentrations of the dye in the dye-bath and 
in the fibre do not stand to each other in a relation 
of simple proportionality, but the concentration in the bath 
is roughly proportional to a power (usually 3 to 5) of the 
concentration in the fibre. Now on the theory of solid 
solutions this indicates that the molecule of the dye in the 
water is three to five times as great as the molecule of the 
dye in the silk ; but this cannot be the case, for the mole- 
cule of the dye-stuff in aqueous solution can be shown by 
other means to be the simplest possible. The numbers 
rather indicate analogy to the process known as absorption 
from solution. Substances like animal charcoal and 
platinum black have the property of condensing gases in 
the extensive surface they present. Similarly they can 
abstract certain substances from solution, as may be seen in 
the employment of animal charcoal for the decoloration of 
solutions. The relation between the concentration in the 
solution and that in the charcoal proves to be of the same 
kind as is met with in dyeing, so that we are led to suspect 
a similarity in the nature of the two processes. The so- 
called "iodide of starch," the blue compound formed when 
starch and iodine solution are brought into contact, would 


appear to be a substance of the same nature as a dyed fibre 
and as charcoal saturated with an acid from solution, for the 
concentrations of the iodine in the aqueous solution and in 
the starch obey approximately the same law as in the other 

We are therefore forced to conclude that whatever 
success has attended the application of the theory of 
solid solutions to other processes, the theory can scarcely 
without modification be accepted as giving an explanation 
of the process of dyeing. 


J. H. VAN'T Hoff. Zeitschrift fiir physikalische Chemie, v., 322 

A. VAN BlJLERT. Ibid., viii., 343 (1891). 
O. N. WlTT. Fdrber-Zeitung, i. (1 890-91). 
C. T. HEYCOCK and F. H. Neville. Journal of the Chemical 

Society, Ixi., 888 (1892). 

E. A. Schneider. Zeitschrift fiir physikalische Chemie, x., 425 

A. Ferratini and F. GARELLI. Gazzetta chimica italiana, xxii., 

ii., 245 ; xxiii., i., 442 (1893). 

F. GARELLI. Ibid., xxiii., ii., 354 (1893). 

F. GARELLI. Zeitschrift fiir physikalische Chemie, xviii., 51 (1895). 
F. W. KtJSTER. Ibid., xii., 508 (1893); xiii., 445 (1894) 5 xv »-> 357 

F. W. KtJSTER. Liebigs Annalen, cclxxxiii., 360 (1894). 

C. HoiTSEMA. Zeitschrift fiir physikalische Chemie, xvii., 1 

G. C. Schmidt. Ibid., xv., 56 (1894). 

E. Beckmann and A. Stock. Ibid., xvii., 120 (1895). 
J. M. VAN BEMMELEN. Ibid., xviii., 331 (1895). 
G. V. GEORGEVICS. Monatshefte fiir Chemie, xv., 705 (1894). 
G. v. Georgevics and E. Lowy. Ibid., xvi., 345 (1895). 

James Walker. 




IN the study of the histological anatomy of plants, apart 
from the structure of the individual cell, the greatest 
advances of the last two decades have been made rather by 
the establishment of new points of view than by the dis- 
covery of new facts. Twenty years ago the solid founda- 
tions of the subject had been securely laid, and a consider- 
able portion of the imposing fabric of histological detail 
which now rests upon them had already been built up. 
This fact is most clearly brought out by the masterly 
summary of existing anatomical knowledge published by 
De Bary in 1877. But splendid monument as it is of its 
author's unsurpassed knowledge of his subject, there can be 
few who have not felt that the Vergleichende Anatomie 
is, as a whole, essentially unreadable. Compare it, in 
imagination, with Sachs' Vorlesungen or with Haber- 
landt's Physiologische Pflaiizenanatomie , and we are 
forced to recognise that De Bary's work is rather an ency- 
clopaedia than a piece of great scientific literature. The 
cause is to be found in the simple fact that there did 
not exist in 1877 a philosophy of the morphological aspect of 
the subject capable of informing " an epitome of the pre- 
sent knowledge of 'the Anatomy of the Vegetative Organs 
of Vascular Plants,' " as the idea of adaptation informed the 
works of Sachs and Haberlandt. 

It is nothing less than the establishment of such a 
philosophy that we now owe to the great Frenchman, Van 
Tieghem. The most important part of his ideas is con- 
tained in what we may call the Stelar Doctrine of Vascular 
Tissue, and it is with this that we shall here be exclusively 

Although the foundations of the stelar theory were laid 
many years ago, outside France it has made its way very 
slowly. In Germany even now it is apparently ignored 


by the majority of anatomists, notwithstanding its accept- 
ance by the most brilliant of German contemporary in- 

In England, though these ideas have recently been 
made familiar to the student by more than one of our lead- 
ing botanists, their discussion has still the interest of com- 
parative novelty. And although the general idea of the 
stele as a morphological unit is simplicity itself, yet the 
application of this idea is in some cases by no means easy, 
so that not only does Strasburger's interpretation of certain 
facts differ from Van Tieghem's, but the author of the 
theory has himself been led to modify his original views in 
an important manner. The possibility of such a difference 
in the interpretation of facts which are undisputed seems to 
spring, if we may say so without presumption, from a 
certain want of definiteness in the apprehension of the 
criteria legitimate to their interpretation. 

To investigate these criteria and to endeavour to as- 
certain their relative validity is one of the primary objects 
of the present paper. 

We shall begin with an account of the development of 
the stelar doctrine. 


In 1 8 70- 1, Van Tieghem published, in the Annales des 
Sciences Nattirelles, a memoir (1) which was to have been 
the first of a series entitled " Recherches sur la syme.trie 
de structure des plantes vasculaires". This instalment con- 
sisted of a general introduction setting forth the plan of 
the whole work, followed by 274 pages devoted to an ex- 
tended anatomical account of the root, in vascular plants. 

The introduction is of the Greatest interest. The 
author tells us how he wished to obtain anatomical defini- 
tions of root, stem, and leaf, in order to give a basis to the 
study of comparative anatomy. These definitions are to 
be framed in accordance with the different kinds of sym- 
metry exhibited in the arrangement of the vascular strands 
in the three organs, to each of which a separate memoir 


is to be devoted. The results so obtained are to be 
applied, in a further series of memoirs, to the solution 
of a number of morphological problems, such as the true 
nature of tendrils, tubers, spines, phylloclades, ovules, etc., 
and finally, to the elucidation of the laws of symmetry 
governing the structure and relations of the ideal colony 
that would be formed if every seed germinated in situ. 

This elaborate scheme for "un cercle d'etudes an- 
atomiques complets et fermes " enables us to understand 
the strength and the weakness of the author's stelar theory. 
The imperative desire to reduce the anatomy of vascular 
plants to a perfect system depending upon simple laws of 
symmetry governing the arrangement of the vascular tissue, 
has been the means of giving us a doctrine, luminous indeed, 
and of wide significance, but scarcely of that rigidly uni- 
versal application which its author claims. But here again, 
as is so often the case in the history of science, the attempt 
to work out logically the various implications of such a 
theory, has been of the utmost value in clearing our ideas 
and extending knowledge, not only by stimulating to the 
discovery of new facts, but by forcing us to examine the 
foundation of our conceptions. 

Of Van Tieghem's scheme, as it stood in 1870, how- 
ever, only the first memoir, that on the root, was ever 
written. The author demonstrates the fundamental identity 
of structure in the roots of all vascular plants, and obtains his 
anatomical definition based on the symmetry of the vascular 
system. He shows that the vascular tissue of a young root 
forms a central cylinder which contains near its periphery 
" faisceaux liberiens" (phloems) alternating with "faisceaux 
vasculaires" (xylems) united by "cellules conjonctives ". 
Hence the vascular system is symmetrical in relation to a 
line, which is the organic axis of the organ. The stem 
agrees with the root in this last point, but on the other 
hand has its "faisceaux libero-vasculaires," " reunis directe- 
ment par le parenchyme primordial ". Where the main 
root passes into the main stem there occurs a "cessation du 
tissu conjonctif special, qui se trouve remplace par le 
parenchyme primitif ". This sentence is specially interest- 


ing because it shows that when it was written Van Tieghem 
had no idea of a central cylinder in the stem. 

Two years later, however, in 1872, in describing (2) 
the transition from root to shoot in Tagetes patiila, he 
writes how the " membrane protectrice " (endodermis) is 
continued up into the stem, retaining its characteristic 
thickenings, and immediately internal to it the "membrane 
rhizogene" of the root (later named the pericycle) is also 
found in the stem still giving rise to rows of lateral roots, 
one row arising from each interval between two bundles. 
Opposite the bundles, however, he holds that the "membrane 
rhizogene" is interrupted, since here the endodermis abuts 
directly on the group of fibres capping the bundle, fibres 
which in accordance with the current opinion he considered 
to belong to the phloem. Here then we have the first 
clear description of the continuation of the central cylinder 
of the root into the stem, and the idea of this continuation 
is the fundamental idea of the stelar theory. It is most 
clearly expressed in a note on p. 112, " Ainsi, et j'insiste 
sur ce point, la tige est, comme la racine, et dans toute son 
etendue, composee d'un cylindre central et d'un parenchyme 
cortical limite en dehors par un epiderme, en dedans par 
une membrane protectrice ou endoderme ". 

The generality of this condition is further insisted upon : 
" Le caractere sur lequel je viens d'appeler l'attention se 
retrouve dans la tige de la grande majorite des plantes 
vasculaires, mais il souffre pourtant quelques exceptions. 
M. Caspary a montre, en effet, que dans quelques plantes 
(Minyanthes trifoliata, Adoxa moschatellina, Bi-asema 
peltatd) chaque faisceau constitutif de la tige est indivi- 
duellement entoure par une membrane protectrice a cellules 
plissees (' Bemerkungen liber die Schutzscheide,' in Pring- 
scheims Jahrbucher, 1865-66, iv., p. 10 1). J'ai retrouve le 
meme fait sur quelques autres plantes, notamment sur 
r Hydrocleis Humboldtii. Dans ce cas, il n'y a pas non 
plus de membrane rhizogene dans les entrenceuds de la 
tige, et il n'existe aucune solution de continuity aucune 
distinction reelle entre le parenchyme cortical et la moelle ' 
(p. 113). This paragraph shows clearly that thus early 


Van Tieghem had recognised the condition which he 
afterwards described as " astely ". 

The "membrane rhizogene," now considered, under the 
name of pericycle, as forming merely the external layer of 
the conjunctive tissue of the cylinder, was at that time 
treated as a region external to, and distinct from, the rest 
of the parenchyma, to which the name " conjonctif " was 
given. But the clear recognition of the existence of an 
individualised stem cylinder, forming a direct continuation, 
tissue for tissue of that of the root, was the first and funda- 
mental step in the evolution of the stelar idea. 

Little progress was made during the next ten years in 
the development of this conception. 

Falkenberg (3), in 1876, showed that the " Aussen- 
scheide " in monocotyledonous rhizomes corresponds with 
the " pericambium " in roots, both in position and role ; 
and Mangin (4) in 1882 entirely confirmed his results and 
showed that not only adventitious roots but also the 
" reseau radicifere " arises from this layer, which he calls 
the " couche dictyogene ". 

In 1882 Van Tieghem published a short paper (5) in 
which, a propos of the Cucurbitacese, he gives conclusive 
reasons, based upon grounds of comparative anatomy, for 
regarding the fibres in the stem, hitherto called primary 
"bast fibres," as really belonging to the "membrane 
rhizogene ". With these extended limits, this layer forms a 
complete investment of the stem cylinder, just as the peri- 
cambium does of the root cylinder. Since the one layer is 
the direct continuation of the other, and the two correspond 
very largely in function as well as in position, it is clearly 
desirable that they should have a common name. For this 
purpose Van Tieghem introduced the word pericycle, which 
was to supersede the various terms "pericambium," 
" Aussenscheide," "membrane rhizogene," "couche dictyo- 
gene," etc., applied by various writers to the same layer in 
various plants and parts of plants, according to its various 
histological characters and functions. The importance of 
this introduction of the conception of the pericycle was of 
course very great, since it fixes more accurately the external 


limit of the cylinder, and thus brings into greater promi- 
nence the idea, already clearly stated in 1872, of an indi- 
vidualised stem cylinder in direct continuity with that of 
the root. 

The term has eventually, though very slowly, found 
its way into general use. 

In 1884 Morot, a pupil of Van Tieghem, published the 
results of a research (6) devoted to a comparative investiga- 
tion of the pericycle in both root and shoot. 

The publication of Morot's paper brings to an end what 
we may call the first phase in the development of the stelar 


The second phase was inaugurated in an investigation (70) 
by Van Tieghem and his pupil Douliot, of the anatomy of the 
stem of various species belonging to the genus Primula. Their 
observations were carried out on a number of new species from 
the East, as well as on many old species, making together a 
total of 114. They resulted [yd) in a division of the aggre- 
gate genus Primula L., into two segregates, Primula Tourn. 
and Auricula Tourn., as had been already done by Tourne- 
fort, but now based on a fundamental difference in the 
structure of the stem of the two segregate genera. While 
the stem of the species belonging to Primula possesses 
a single normal central cylinder in its whole extent, the 
narrow cylinder of the hypocotyl of an Auricula, instead 
of dilating in the ordinary way above the level of the coty- 
ledons, gives rise by successive bifurcations to two or more 
vascular strands, each surrounded by an endodermis and 
possessing" the structure of the single hypocotyledonary 
cylinder. In the genus Gunnera {Haloragece) a similar 
state of things obtains. These facts were, in the main, 
already known, having been investigated by Vaupell, 
Kamienski and Reinke. The opinion of these authors 
was, however, that the separate vascular strands were vas- 
cular bundles of the " concentric " type with peripheral 
phloem, comparable for instance to those found in the 
cortex of certain Mclastoniace<z ; and this was the view 


taken by De Bary in his classical Vergleichende Anatotnie. 
The numerous vascular strands in the rhizomes of most 
Leptosporangiate Ferns were regarded by De Bary in the 
same light. 

But Van Tieghem, having, as we have seen, come 
to regard the central cylinder rather than the bundle as 
the morphological unit of vascular tissue in both root and 
shoot, was now led to the conclusion that in Auricula, 
Gunnera and the majority of Ferns 1 we have really to 
deal with a splitting- of the single cylinder of the hypocotyl, 
as we trace it upwards, by successive bifurcations, into a 
number of such cylinders (jc and 8). Van Tieghem 
and Douliot proposed to call such a cylinder a stele (Greek 
<7r/;X»7, a column). A root or a stem containing one such 
stele would be monostelic, if it contained more than one 
polystelic. A third case was distinguished. If the cylinder 
of the hypocotyl breaks up, as it is traced upwards, into 
its component bundles, each of which is surrounded by a 
special endodermis, the cylinder, according to our authors, 
no longer exists ; the stem is astelic. This case, already 
described in 1872, obtains in the stems of various Ranun- 
culacece, in Nymphceacece, in Hydrocleis, in some species of 
Equisetum, etc., as well as in the majority of petioles and in 
blades of all leaves. 

Cases of Polystely fall into two groups. First, where 
on a transverse section the various steles are seen to be 
completely separate, we have a state of dialystely. Secondly, 
where the steles are united laterally, so as to form a more or 
less complete ring in transverse section, enclosing a more or 
less isolated portion of extra-stelar tissue, which occupies the 
centre of the ring, we have a state of gamostely. These 
two conditions are not to be sharply separated, since the 
steles of all polystelic stems show more or less frequent 
lateral unions, and the gamostelic condition is simply a 
case where these unions are very frequent and persistent. 

1 Leclerc du Sablon in 1890 (9) worked out the connections, in 
several Ferns, of the single hypocotyledonary cylinder with the cylinders 
of stem. 


We may tabulate the results thus obtained as follows: — 

Monostely. — A single central cylinder. All roots and 
hypocotyls, nearly all Phanerogamic stems, and stems of 
many Vascular Cryptogams. 

Polystely. — More than one cylinder. Stems of most 
ferns, most species of Selaginella, and among Phanerogams 
of Auricula and Gunner a} 

(a) Dialystely.— Steles separate for most of their 
course. Most Ferns Selaginella and. Auricula ursi, etc. 

(6) Gamostely. — Steles united laterally for most of their 
course. Marsilia, Pilularia, Pteris aurita, etc. Auri- 
cula japonica, etc. 

Astely. — No cylinder. Leaf blades, most petioles, 
stems of some species of Equisetum and Ranunculus^ 
stems of Hydrocleis, Ophioglossum, Limnanthonum, Nym- 
phceacece, etc. (yc). 

The publication, in 1890-91, of the second edition of 
Van Tieghem's Trait e" de Botanique (10), which contains a 
full exposition of the stelar doctrine on the lines indicated, 
may be said to mark the close of the second phase in the 
development of the theory. 


The third phase, from that date to the present time, has 
been occupied by various developments and modifications of 
the doctrine on the part of the author and his pupils, and has 
been marked by considerable criticism, mainly of these 
newer developments. 

The first line of research that calls for notice is a re- 
investigation of the conjunctive tissue of the typical central 
cylinder of the flowering plant. This has led Flot (11) to 

1 In a paper recently communicated to the Linnean Society, Mr. B. 
G. Cormack describes cases of polystely met with in the adventitious roots 
of three genera of Palms, viz., Areca, Cocos and Verschaffeltia. It appears 
that the single stele of the root splits, as it is traced downwards, into a ring 
of separate steles. Later on these steles again pass over into a single 
cylinder. This seems to be an important modification of Polystely as 
described by Van Tieghem and Douliot, and Leclerc du Sablon. 


add a new region to those already distinguished. He 
finds a zone situated at the periphery of the pith, i.e., just 
internal to the ring of bundles, corresponding exactly to 
the pericycle external to the ring, as well characterised 
histologically as the pericycle itself, and indeed resembling 
the latter very closely in structure and role. This zone, 
the perimedullary zone, is according to Flot (and his 
figures entirely support this) separate in development from 
the pith proper, or internal conjunctive, and belongs rather 
to the hollow cylinder of tissue (the "thickening ring " of the 
older German anatomists) giving rise to the bundles and 
the conjunctive immediately surrounding them {external 
conjunctive). It is impossible sharply to separate the peri- 
medullary zone on the one side, just as Morot found it 
impossible to separate the pericycle on the other, from the 
ray tissue, and we should rather regard the contrast of 
the pith with the external conjunctive tissue, as of greater 
importance than the division of the latter into pericycle, 
rays and perimedullary zone, which are in the main 
topographical regions marked out by the limits of the 
bundles. In many adult stems it is however impossible to 
fix the limits of external and internal conjunctive, just as 
it is often impossible to fix the limits between external con- 
junctive and cortex. Flot is of opinion that this is owing 
to a growth in breadth of the cells of the external conjunc- 
tive continued longer than in the pith, the whole of the tissue 
of the cylinder thus becoming approximated in size and 
shape. This same cause, together with a masking of the 
endodermal thickenings (in cases where these are originally 
present) by a general thickening of the walls of all the 
parenchyma cells may very conceivably account for the 
frequent absence of the obvious limit between cortex and 
cylinder, though we are not aware that such an occurrence 
has been either established or suggested. 1 Further in- 
vestigation on this point, as well as on the separation of 
the regions in root cylinders with a well-developed con- 

X I now find that Sanio (24, pp. 371-2) states that this is practically 
what occurs in the stem of Ranunculus acris. 


junctive system, is much needed to complete our know- 
ledge of these matters. 

An important modification of the theory of steles has 
been made by Van Tieghem himself in extending the use 
of the term astely so as to make it include the state of 
things obtaining in the stems of all species of Equisetum 
(12), and of p hio gloss ace ce (13). 

Let us take first the case of Equisetum. Well- 
marked endodermes are found in the stems of all species, 
but their disposition, which was fully worked out many 
years ago by Pfitzer, is very various, not only in different 
species, but in different parts of the stem of the same 
species. There are three types of arrangement. In the 
first each vascular bundle is surrounded by a special endo- 
dermis ; in the second the ring of bundles is bordered within 
and without by a general endodermis ; and in the third 
the outer endodermis alone is present. In the second 
edition of the Traite Van Tieghem assigned the first 
two conditions to the astelic, the third to the monostelic 
type, but in a paper (12) published in the same year (1890) 
he calls attention to the fact, discovered by Pfitzer, that 
all the species possess, at their nodes, the first or second of 
the arrangements in question. He therefore concludes 
that all belong really to the astelic type, and that where, 
for instance, the second type, just above a node, passes 
back into the third, we have simply a case of the dis- 
appearance of the special characters of the inner endo- 
dermis, which must still be supposed to exist. The " mono- 
stely" is only apparent, and the tissue bordering the 
central canal of the stem, internal to the inner (theo- 
retical) "endodermis," is not in reality pith, but rather 
" inner cortex" (extra-stelar tissue). The first of the three 
arrangements is to be called dialydesmic, since each bundle 
with its sheath of conjunctive is separate ; the second and 
third gamodesmic, since the conjunctive tissue surrounding 
the bundles is in lateral confluence. 

Turning now to the Ophioglossacece we have a similar 
argument (13). The stem of Ophioglossum vulgahtm, 
below the level of the first leaf, is monostelic, but above the 


first leaf contains five separate bundles each with a separate, 
though feebly suberised, endodermis. Hence it was 
treated by Van Tieghem, in the Traits, as astelic. In 
Botrychium Lunaria, whose stem is also monostelic at the 
base, the endodermis, after the departure of the first leaf 
trace, does not close round each separate bundle but 
becomes as it were invaginated into the cylinder, so that 
the vascular tissue forms on transverse section a horse- 
shoe bounded by the endodermis. The free edges of the 
horseshoe meet, as we pass up the stem, and the inner 
portion of the endodermis becomes entirely separated from 
the outer, so that we have an equivalent of the second or 
gamodesmic condition found in the stems of Eqtiiseta. 
Higher up the inner endodermis loses its thickenings, just 
as in some Equiseta, and this gives us an apparently 
monostelic condition. In accordance with his revised 
view, Van Tieghem considers that OpJiioglossum has an 
astelic-dialydesmic stem, while those of Botrychium and 
Helminthostachys are astelic-gamodesmic. 


It will be most convenient to introduce here a critical 
investigation of the stelar theory as thus modified by its 
author, and so far as it depends upon the morphological 
interpretation of the arrangement and relations of vascular 
tissue in the adult organs of vascular plants ; deferring for 
the present a consideration of the developmental facts 
bearing upon the theory. 

There is no need to discuss at any length the funda- 
mental conception of the stele arrived at in the period which 
we have called the first phase of development of the idea. 
It depends upon the tracing into the stem of the root 
cylinder, and upon the demonstration that its characters as 
a cylinder are maintained in the latter. This demonstration, 
begun, as we have seen, in 1872, eventually led to the 
explicit recognition of the fact that the system of bundles 
forming the central cylinder possesses morphological charac- 
ters much more constant than those of the vascular bundle, 


and is hence more worthy to be taken as the morphological 
unit of vascular tissue. It is indeed impossible to give a 
morphological definition of a vascular bundle at all. " From 
the very first those bundles which consist essentially of 
definitely arranged groups of tracheae and sieve tubes . . . 
have been called vascular bundles" (14, p. 232, Eng. ed.). 
But thus defined, a "vascular bundle" has no constant 
histological characters beyond the fact of containing both 
xylem and phloem. According to the arrangement of these, 
bundles have been classified as radial, concentric, collateral, 
etc. Such an arrangement brings together vascular strands 
of very different orders of complexity. In the first place 
it associates the axial cylinder ("radial bundle") of a root, 
possessing a number of quite distinct xylem and phloem 
strands, with the "collateral bundle" of a Phanerogamic 
stem, formed of a single strand of xylem and phloem in close 
association, the latter being continuous moreover with a 
portion only of the former. Again it associates even more 
closely under the term " concentric bundle " the vascular 
strands found in the stem of Auricula, Gunnera and Ferns 
with those of quite different structure found in the pith and 
cortex of Melastomacecz, etc. 

Such a classification is clearly, from a morphological point 
of view, quite artificial. But if we extend the use of the term 
bundle, as is often done, so as to include strands of tracheae 
alone, and of sieve tubes alone, we can retain it as a con- 
venient word without morphological connotation, and 
applicable to any strand of tissue belonging to the vascular 
system. And we may then qualify the word by any adjec- 
tive we choose without morphological implication. Thus we 
may speak of the composite radial bundle of the root as 
composed of separate xylem bundles and phloem bundles 
alternating at its periphery ; of the concentric bundle of the 
stem of an aquatic plant as sometimes composed of separate 
collateral bundles, in other cases consisting simply of a con- 
tinuous cylinder of phloem surrounding a central strand of 
xylem ; of the concentric bundle of a fern petiole gradually 
passing to the collateral type as we trace it into the lamina, 
and so on. Meanwhile the study of the homologies of the 


various strands is quite a distinct matter, and requires a 
distinct terminology. 


The acceptance of the central cylinder in the " mono- 
stelic " stem as a region of the first morphological rank is 
now very general. The only criticism which we have to 
consider is that which calls attention to the frequent want 
of definiteness about its external limit, and is inclined on 
this ground to question its individuality. This want of 
definiteness arises from the absence, in many adult stems, 
of the special characters of the endodermis (innermost 
layer of the cortex), often combined with an identity in 
size, shape and characters of cell-membrane between 
the cells of the cortex and those of the conjunctive. Such 
a state of things obtains, to take a single instance, in the 
stem of Ranunculus repens. A transverse section of such 
a stem shows the separate bundles imbedded in a homo- 
geneous ground tissue, and to speak of a well-marked central 
cylinder is to speak of that which does not, in fact, exist. 

Now this, as it stands, is a perfectly legitimate criticism, 

and its force as against the general validity of the stelar 

idea depends simply upon the greater or less generality of 

the condition described. Van Tieghem (10, p. 752) states 

that when, after the formation of the endodermis, the stem 

undergoes considerable intercalary growth, the folds on 

the radial walls of the endodermal cells become stretched out 

so that they become difficult or impossible to see. In other 

cases no suberisation of the radial walls occurs, and then, 

unless the endodermal cells are distinguished by possessing 

starch, it is admitted that the limit of the cylinder is difficult 

to determine, but says Van Tieghem {Joe. eit.) : " il reste la 

forme differente des cellules ". This, however, as has been 

said, is by no means always obvious. A possible cause of 

such a condition, assuming the limits of the young cylinder 

to be well defined, has already been suggested, but 

new investigations are necessary to determine the point. 

If, for the sake of argument, we make the opposite assump- 



tion, that the vascular bundles are sometimes differentiated 
in the middle of a homogeneous ground tissue, no trace 
of a special endodermis or pericycle being visible at any 
time, we could not predicate the existence, in such cases, 
of a central cylinder in the stem. And further, if such 
a condition obtained in the majority of instances (certainly 
an unlikely supposition) we should not, of course, be 
justified in predicating the general existence in the stem 
of a central cylinder, and this would necessitate such a 
radical modification in the generalised statement of the 
facts, that the stelar idea would lose the greater part of 
its significance. We shall have to recur to a discussion 
of the limit of the cylinder, but these simple considera- 
tions are insisted upon here, because they are apparently 
lost sight of in much of the current writing of Van Tieg- 
hem's adherents. It seems to be implicitly assumed that if 
a good anatomical distinction can be made in a certain 
number of cases, it is permissible to generalise the distinc- 
tion and erect it into a morphological doctrine. The 
existence of those cases to which the doctrine does not 
apply is either ignored, or the distinction is said to be 
" theoretical ". There is of course no such thing as a true 
"theoretical" distinction which is not also actual. The 
fallacy arises from a tendency to regard all morphological 
doctrine as of absolute value, whereas its value is never 
anything but relative. What we have to decide in any 
given case is the amount of this relative value, and whether 
that amount is sufficient to make the doctrine express a 
general truth so far as the objects under consideration are 

The foregoing reflections lead us naturally to consider 
those cases which Van Tieghem himself excepts from the 
application of the stelar doctrine, namely, the cases of 
"astely". Already in the earliest paper containing the 
germ of the stelar idea we find certain cases not covered 
by the general statement of the existence of a cylinder in 
the stem. In 1886 these cases together with other similar 
ones were called astelic, and more recently still the concep- 
tion has been further elaborated. 


The conception is governed throughout by the idea of 
the endodermis as a definite morphological layer, always 
separating" stelar from extra-stelar tissue. And the en- 
dodermis is to be recognised by the suberised thicken- 
ings on its radial walls. It is simply by the disposition of 
layers of cells so thickened that we are supposed to be 
able to distinguish the various arrangements described. 
It is easy to show that this criterion is quite illegitimate. 
The term endodermis is defined by Van Tieghem as the 
innermost layer of the cortex which " offre frequemment " 
the special character in question (10, pp. 738-9). Not 
only, however, do cell layers with the same character 
occur in quite other situations {e.g., in the middle of many 
periderms), but the innermost layer of the cortex certainly 
does not always possess it. So that these thickenings 
cannot be used to mark a layer of invariable morpho- 
logical value. And even in Equisetum, Van Tieghem does 
not keep to his own criterion. For when the " astelic 
gamodesmic " passes to the apparently monostelic con- 
dition we are told that the inner endodermis is still 
present though its special characters have disappeared. 
But, we may well ask, if such great importance is to be 
attached to these special characters as to justify us in 
founding new types of structure simply upon the disposition 
of the layers exhibiting them, why should we be suddenly 
asked to recognise as equivalent a layer which does not 
exhibit them ? The criterion becomes completely chimeri- 

Strasburger (15) has pointed out that an endodermoid 
layer is an air-tight barrier which does not prevent the 
passage of water through its cells. Such a layer is found 
in a position to shut off the water-conducting system of a 
plant from its air-containing lacunar system, but this posi- 
tion may vary within the same genus [Ranunculus, 
Equisetum), and has no necessary connection with any 
morphological region. As a matter of fact it is most often 
formed from the inner layer of the cortex, but may be 
developed from conjunctive tissue, or even (leaf of Isoetes) 
from intra-fascicular parenchyma. Since the innermost 


layer of the cortex does not always possess the special 
thickenings which give it the right to be called a 
" membrane," Strasburger objects to Van Tieghem's re- 
definition of the word endodermis, and proposes to sub- 
stitute the term Phloeoterma, to be applied to the inner 
layer of the cortex, i.e., to be used in the strictly morpho- 
logical sense, whether this inner layer has special characters 
or not, and to reserve the term endodermis in accordance 
with its original sense for any sheath or membrane com- 
posed of cells with suberised radial walls or other dis- 
tinctive thickenings, without reference to its position. 
This revised terminology certainly helps us to get rid 
of the confusion of thought manifested in Van Tieg- 
hem's use of the word endodermis. Strasburger concludes 
that as all species of Equisetum agree in possessing a ring 
of simple collateral bundles, they should all be considered 
monostelic, whether the phloeoterma be developed as a 
general endodermis, or each bundle possess a special 
endodermis, the phloeoterma having no characters by which 
it can be distinguished. The same considerations would 
apply to the genus Ranunculus and the other cases of 
"astely ". While we must fully admit the general force of 
his argument on the ground of comparative anatomy, it is 
difficult to agree with the following sentence : " Die 
Grenze der Rinde gegen den Centralcylinder ist dort wo 
sie sich nicht besonders als Endodermis oder Starkescheide 
markirt, nur theoretisch zu ziehen, dieselbe ist aber flir alle 
Falle festzuhalten ' : (15, p. 484). How is one to "hold 
fast " a limit which one cannot distinguish ? We can only 
refer to the remarks which have been already made upon 
this subject, but we shall briefly recur to the subject in 
considering the development of the stele. 

Leaves furnish us with excellent examples of the frequent 
impossibility of separating stelar from extra-stelar tissue. 
Putting aside those cases in which one or more steles 
from the polystelic stem directly enter the petiole 
[Gunncra, Ferns), we have to consider the ordinary 
case in a flowering plant, where we have one or more 
bundles leaving the cylinder and passing into the petiole. 


These bundles are accompanied by a certain amount of 
closely associated parenchyma belonging to the external 
conjunctive of Flot, a tissue which in the leaf Van Tieghem 
now calls peridesm (16). The bundles are sometimes 
arranged in a ring, and the whole may be, though com- 
paratively rarely, surrounded by an endodermis. The 
petiole is then, according to Van Tieghem (10, p. 842), 
monostelic. In the commoner case where each bundle has 
an endodermis of its own the petiole is astelic. 

Strasburger prefers the term schizostelic (15), since the 
stelar tissue of the petiole represents a separated portion or 
portions of that of the stem. To such a portion he gives the 
name schizostele or schistostele \ at the same time denying the 
existence of monostelic petioles in Phanerogams on the 
ground that the apparent pith of the petiole is continuous 
with the cortex, and not with the pith, of the stem. This 
last contention brings forward a difficult position. Is it de- 
sirable to introduce the question of continuity at all ? If we 
have in the petiole a structure apparently identical with that 
which we have agreed to call monostelic in the stem, should 
we be satisfied to call it monostelic here also, without con- 
sidering the connections of its parts with those of the stem ? 
The strength of Strasburger's position lies in the fact that 
the continuity, region for region, of the cylinder of root 
and stem is really the basis of the stelar idea. The origin 
of the difficulty is to be found in the tendency of a petiole, 
where it is subject to the same conditions as a stem, to 
assume the characters of a stem, and among them the 
arrangement of its vascular tissue according to a radially 
symmetrical type. We might, perhaps, fitly call such a 
structure a pseudostele. 

The mesophyll of the leaf (corresponding with the cortex 
of the stem) which surrounds the smaller vascular bundles, 
often has its innermost layer or phloeoterma, which abuts 

1 Van Tieghem has since (17, p. 285) used the word meristele for 
Strasburger's " schizostele," and applied the latter term to the portion of 
stelar tissue enclosed by each special endodermis in an "astelic" stem. 
This seems an unwarrantable diversion of the meaning of Strasburger's 


immediately upon the peridesm of the bundle, specially 
characterised. The cells of the phloeoterma are often de- 
prived of chlorophyll, or this is confined to the side walls, 
and these walls may also be suberised. It is, however, a 
rare case for such layers to be united in a continuous 
system with the phloeoterma of the stem, and thus to shut 
off completely, by means of a continuous membrane, the 
entire stelar system of the plant from its cortical tissue. 
This state of things obtains, however, in Pinus and some 
dicotyledonous genera, e.g., Galium. In most dicotyledo- 
nous petioles endodermoid layers, if distinguishable at all, 
are often incomplete and not necessarily formed from the 
phloeoterma. A closed sheath to the bundles is, however, 
often formed in Angiosperm petioles by thickened peri- 
desmic (stelar) tissue, such a sheath being called by Stras- 
burger a stelolemma (15). The ensemble of the phenomena 
shows us, clearly enough, that the endodermis, in its original 
sense, cannot be taken here, any more than in the stem, 
as a layer of constant morphological value. The phloeo- 
terma may be distinguishable by endodermal or other char- 
acters, but on the other hand, it may not. 

The main fact in regard to the vascular system of the 
leaf is one which was pointed out by Van Tieghem in 1870. 
The system is bilaterally symmetrical in relation to the 
plane including the organic axes of both leaf and stem, 
and not, like that of root and stem, radially symmetrical 
about its organic axis. The designation of the continuous 
cylinder of root and stem as a stele and of each bundle or 
the whole bundle system of the leaf as a schistostele or 
meristele is in complete accord with this general fact. But 
we must not disguise from ourselves that both the stele and 
the meristele may not exist in the adult as sharply separated 

A. G. Tansley. 
( To be contimted. ) 



IN my previous article I gave some account of a research 
by Heidenhain in which this observer, after drawing 
certain deductions from the theory of osmotic pressures, shows 
that the phenomena of absorption from the intestinal canal are 
irreconcilable with these deductions, and are therefore not 
susceptible of a mechanical explanation, but must be as- 
cribed to the active intervention of cells. Since analogous 
problems to those discussed by Heidenhain are continually 
coming before us in physiology, it is important that we 
should have a clear idea of the factors which are involved 
in the passage of water or dissolved substances across 
membranes. I therefore propose to reproduce Heiden- 
hain's statements, and then to consider how far they are 
true for the special cases which occur in the body. 

These statements are as follows : — 

i. If two watery solutions with the same osmotic pres- 
sure are separated by a membrane through which diffusion 
can take place, no change in volume occurs on either side 
of the membrane. 

2. If the solutions on either side of the membrane are of 
unequal osmotic pressure, water passes from the side where 
the pressure is less to the side where the osmotic pressure 
is greater. 

3. The osmotic pressure of a solution is equal to the sum 
of the partial pressures of the various dissolved substances. 

4. If the solutions on the two sides of the membrane 
have the same total osmotic pressure but unequal partial 
pressures of their various constituents, each constituent of 
the solution passes from the side where it has the higher 
partial pressure to the other side. No change in the volume 
of water on the two sides takes place. 

Of these four statements only one (No. 3) is absolutely 


correct. The other three are only correct under certain 
defined conditions which are rarely fulfilled in the body. 
There are factors at work which have been practically dis- 
regarded by most of the recent workers on the subject, and 
which may tend to produce movement of fluid in apparent 
opposition to the difference of osmotic pressure. Instances 
of such cases are afforded in a paper by Lazarus Barlow, to 
a consideration of whose work we shall shortly return. 

There can be no doubt that in the phenomena of trans- 
ference of fluid or dissolved substances across a membrane 
the nature of the membrane itself is all-important. I will, 
therefore, shortly run through the various modes in which 
interchanges may take place across membranes of varying 
permeability. We shall see that the close analogy which 
exists between substances in solution and gases, when 
dealing with "semi-permeable" membranes, is also borne 
out by experiment when used to predict the behaviour of 
solutions separated by such permeable membranes as occur 
in the body. 

The simplest case is that in which two fluids are sepa- 
rated by a perfect semi-permeable membrane that permits 
the passage of water but is absolutely impermeable to dis- 
solved substances. In this case the transference of water 
from one side to the other depends entirely on the difference 
of osmotic pressure between the two sides. 




If we suppose two vessels, A and B, separated by such 
a membrane, A containing a solution of a and B a solution 
of (5, water will pass from A to B so long as the osmotic 
pressure of /3 is greater than the osmotic pressure of the 
solution of a. If B be subjected to a hydrostatic pressure 
greater than the osmotic difference between the two fluids, 
water will pass from B to A until the force causing filtration 
or transudation (the hydrostatic pressure) is equal to the 


force causing absorption into B (the difference of osmotic 
pressures). Under no circumstance will there be any trans- 
ference of salt or dissolved substance between the two sides. 
Such semi-permeable membranes as this, however, rarely 
occur in the body. It is possible that the external layer of 
the cell-protoplasm may in some cases resemble the proto- 
plasmic pellicle of plant-cells in possessing this "semi-per- 
meability " ; but in nearly all cases where we have a mem- 
brane made up of a number of cells, it can be shown that 
such a membrane permits the free passage of at any rate a 
large number of dissolved substances. 

Let us now consider what will occur when the two solu- 
tions A and B are separated by a membrane which permits the 
free passage of salts and water. If the osmotic pressure of 
B be higher than A at the commencement of the experi- 
ment, the force tending to move water from A to B will be 
equal to this osmotic difference. But there is at the same 
time set up a diffusion of the dissolved substances from B 
to A and from A to B. The result of this diffusion must 
be that there is no longer a sudden drop of osmotic pressure 
from B to A, and the result of the primary osmotic difference 
on the movement of water will be minimised in proportion 
to the freedom of diffusion which takes place through the 
membrane. Now let us take a case in which A and B re- 
present equimolecular and isotonic solutions of o and /3. 
It is evident that the movement of water into A will vary 
as Ap - Bp l = O. But diffusion also occurs of a into B and 
of (3 into A. Now the amount of substance diffusing from 
a solution is proportional to the concentration, and there- 
fore to its osmotic pressure, as well as to its diffusion 

Hence the amount of a diffusing into B will vary as 
Aft . ak (when k is the diffusion coefficient). 

In the same way the amount of (3 diffusing into A will 
vary as Bp, (5k'. 

Hence if ak is greater than (3k', i.e., if a is more diffusible 
than (3, the initial result must be that a greater number of 

1 Ap = osmotic pressure of A, etc. 


molecules of o will pass into B than of /3 into A. Hence 
the solutions on the two sides of the membrane will be no 
longer equimolecular, but the total number of molecules of 
a + (3 in B will be greater than the number of molecules of 
a + j3 in A, and this difference will be most marked in the 
layers of fluid nearest the membrane. The result therefore 
of the unequal diffusion of the two substances is to upset 
the previous equality of osmotic pressures. The layer of 
fluid on the B side of the membrane will have an osmotic 
pressure greater than the layer of fluid in immediate contact 
with the A side of the membrane, and there will thus be a 
movement of water from A to B. Hence if we have two 
equimolecular and isotonic solutions of different substances 
separated by a membrane permeable to the dissolved sub- 
stances, there will be an initial movement of fluid towards 
the side of the less diffusible substance. 

We have an exact parallel to this in Graham's familiar 
experiment in which a porous pot filled with hydrogen is 
connected by a vertical tube with mercury. In consequence 
of the more rapid diffusion outwards of the hydrogen than 
of atmospheric air inwards, the pressure within the pot sinks 
below that of the surrounding atmosphere, and the mercury 
rises several inches in the tube. We must therefore con- 
clude that even when the two solutions on either side of the 
membrane are isotonic, there may be a movement of fluid 
from one side to the other with a performance of work in 
the process. 

The experimental proof of the truth of this argument is 
to be found in a recent paper by Dr. Lazarus Barlow. 
This observer — after pointing out that the huge total 
osmotic pressures of the salt solutions in the body can very 
seldom come into play — insists on the fact that the most im- 
portant point to study in this regard is the initial changes 
that take place between dissimilar fluids separated by a 
membrane — as he terms it — the initial rate of osmosis. For 
this purpose he employs a funnel, the neck of which is pro- 
longed into a capillary tube, while on the mouth is tied a 
piece of peritoneal membrane. The funnel is filled with 
the solution whose osmotic attraction for water it is wished 


to measure, and its mouth covered with the membrane is 
immersed in distilled water or in dilute serum. 

The experiments which are the most interesting are 
those in which decinormal solutions of glucose, urea, 
sodium chloride were compared as to their initial rates of 
osmosis, the outer fluid being water. He concludes from 
his experiments that, in the case of prepared peritoneal 
membrane, the initial rates of osmosis of glucose, sodium 
chloride and urea in equimolecular solutions do not corre- 
spond to the ratio between their final osmotic pressures (as 
estimated by the depression of freezing-point), but the 
initial rate of osmosis of glucose {i.e., the rate with which 
water passes into this solution) is greater than that of 
sodium chloride, and the initial rate of osmosis of sodium 
chloride greater than that of urea. 

In these experiments the only two solutions which 
are strictly comparable are those of urea and glucose 
(A = 0*189° C), since the decinormal Na CI solution had 
nearly double the osmotic pressure of these two (A = 0*35 1). 
In three typical experiments, each of which lasted three 
hours, the average rates at which the fluid in the funnel 
increased in volume during the first hour were : in the case 
of glucose, 7! mm. in five minutes ; in the case of sodium 
chloride, 43 mm. ; and in the case of urea, iJ T mm. 
These figures are evidently not proportional to the differ- 
ence of osmotic pressures between the fluid and the funnel 
and the water in the reservoir. But we have already seen 
that the moving force is not the total difference of pressure 
between the fluids in the vessels on either side of the 
membrane, but the difference of pressure between the 
layers of fluid in immediate contact with each side of the 
membrane. The fall of osmotic pressure across the thick- 
ness of the membrane varies inversely as the rate of 
diffusion of the dissolved substance. The question arises 
therefore whether the results obtained by Lazarus Barlow 
can be accounted for by differences in the rate of diffusion. 
In the carefully worked-out tables by this observer we have 
all the data necessary to decide the question. In the case 
of glucose, the freezing-point of the solution at the begin- 


ning of the experiment was - o'i8q° ; at the end of the three 
hours' experiment it was — ot JJ° C. — corresponding to a 
loss of 6 per cent, of the dextrose. In the case of the 
urea, the freezing-point at the beginning was "189°, and at 
the end was- 0*154° C, a loss of 18 per cent. Here then 
the initial rate of osmosis of the glucose was about five 
times that of the urea ; the loss by diffusion of the glucose 
was about one-third that of the urea. In the case of the 
sodium chloride the loss amounted to 22 per cent. ; but 
here the total difference of osmotic pressure was very 
nearly double that of the other two solutions, and the result 
is that the initial rate of osmosis of the sodium chloride takes 
an intermediate place between that of urea and that of 

In this paper the results of another experiment are 
given to show that osmosis may occur from a fluid having 
a higher final osmotic pressure towards a fluid having a 
lower final osmotic pressure. If, for example, equimolecu- 
lar solutions of sodium chloride and glucose be separated 
by a peritoneal membrane, the osmotic flow will take place 
from the fluid having the higher final osmotic pressure — 
sodium chloride. We might compare with this experiment 
the results of separating hydrogen at one atmosphere's 
pressure from oxygen at two atmospheres' pressure by 
means of a plate of graphite. In this case the initial result 
will be a still further increase of pressure on the oxygen 
side of the diaphragm — a movement of gas against pres- 
sure taking place in consequence of the greater diffusion 
velocity of hydrogen. 

So far we have only considered the behaviour of solu- 
tions when separated by a membrane, the permeability of 
which to salts is comparable to that of water ; so that the 
passage of salts through the membrane depends merely on 
the diffusion rates of the salts. There can be no doubt, 
however, that we might get analogous movements of fluid 
against total osmotic pressure determined, not by the 
diffusibility of the salts, but by the permeability of the mem- 
brane for the salts — a permeability which may depend on a 
state of solution or attraction existing between membrane 


and salts. We have a familiar analogue to such a condition 
of things in the passage of gases through an india-rubber 
sheet. If two bottles, one containing carbonic acid, the 
other hydrogen, be separated by a sheet of india-rubber, 
C0 3 passes into the hydrogen bottle more quickly than 
hydrogen can pass out into the C0 2 bottle, so that a dif- 
ference of pressure is created between the two bottles, and 
the rubber bulges into the C0 2 bottle. We might, in the 
same way, conceive of a membrane which permitted the 
passage of dextrose more easily than that of urea. With 
such a membrane, experiments conducted in the same way 
as Dr. Barlow's, would lead to diametrically opposite re- 
sults. The importance of the membrane in determining 
the direction of the osmotic passage of fluid is well illustrated 
by Raoult's experiments. When alcohol and ether were 
separated by an animal membrane, alcohol passed into the 
ether, whereas if vulcanite were employed for the dia- 
phragm, the osmotic flow was in the reverse direction, 
and an enormous pressure was set up on the alcohol side of 
the diaphragm. 

Here we have a possible clue to the "explanation" of 
many phenomena of cell activity, to which the term " vital" 
is often assigned. In the swimming-bladder of fishes, for 
instance, we find a gas which is extremely rich in oxygen, 
and the oxygen is said to have been secreted by the cells 
lining the bladder. It is, however, very possible that the 
processes here may be exactly analogous to Graham's 
atmolysis, and that the bladder may represent a perfected 
form of Graham's india-rubber bag. 

The next point to be considered is the passage of a 
dissolved substance across membranes in consequence of 
differences in the partial pressure of the substance in ques- 
tion on the two sides of the membrane. Great stress is 
laid by Heidenhain and his pupil Orlow on the fact that 
in the peritoneal cavity, as well as from the intestine, salt 
may be taken up from fluids containing a smaller percentage 
of this substance than does the blood plasma, and they 
regard this absorption as pointing indubitably to an active 
intervention of living cells in the process. This argument 



requires examination. Supposing the two vessels A and B 
to be separated by a membrane which offers free passage 
to water, and a difficult passage to salts. Let A contain '5 




per cent, salt solution and B a solution isotonic with a 1 
per cent. Na CI, but containing only '65 per cent, of this 
salt, the rest of its osmotic tension being due to other dis- 
solved substances. If the membrane were absolutely " semi- 
permeable," water would pass from A to B until the two 
fluids were isotonic, i.e., until A contained 1 per cent. Na CI 
(we may regard volume of B as infinitely great to 
simplify the argument). If, however, the membrane per- 
mitted passage of salt, the course of events might be as 
follows : At first water would pass out of A, and salt would 
diffuse in until the percentage of Na CI in A was equal 
to that in B. There would now be an equal partial pres- 
sure of Na CI on the two sides of the membrane, but the 
total osmotic pressure of B would still be higher than A. 
Water would therefore still continue to pass from A to B 
more rapidly than the other ingredients of B could pass 
into A. As soon, however, as more water passed only 
from A, the percentage of N a CI in A would be raised 
above that in B. The extent to which this occurs will 
depend on the impermeability of the membrane. As soon, 
however, as the Na CI in A reaches a certain concentration 
it will pass over into B, and this will goon until equilibrium 
is established between A and B. Extending this argument 
to the conditions obtaining in the living body, we may con- 
clude that neither the raising of the percentage of a salt 
in any fluid above that of the same salt in the plasma, nor 
the passage of a salt from a hypotonic fluid into the blood 
plasma, can afford in itself any proof of an active interven- 
tion of cells in the process. 


Thus in the case of the pleura we seem to have a mem- 
brane which is very imperfectly semi-permeable. It is per- 
meable to salts, but presents rather more resistance to their 
passage than to the passage of water. Hence on injecting 
•5 per cent. Na CI solution into the pleural cavity water 
passes from the pleural fluid into the blood, until the per- 
centage of sodium chloride in the fluid is raised perceptibly 
above that in the blood plasma. The limit of the resistance 
of the pleural membrane to the passage of salt is, however, 
soon reached, and then salt passes from pleural fluid into 
blood ; but in every case this passage is from a region of 
higher to a region of lower partial pressure. Hence at 
a certain stage of the experiment we find a higher percentage 
of salt in the pleura than in the blood-vessels, although 
the total amount of salt in the pleural fluid is less than 
that originally put in, or, in other words, salt has been 

We have already seen that the effective osmotic pressure 
of a substance, i.e., its power of attracting water across a 
membrane, varies inversely as its diffusibility, or as the 
permeability of the membrane to it. What then will be 
the effect supposing that on one side of the membrane we 
place some substance in solution to which the membrane 
is impermeable ? 

We will suppose that A and B both contain 1 per cent. 
Na CI, but that B contains in addition some substance x to 
which the membrane is impermeable. Since the osmotic 
pressure of B is higher, by the partial pressure of x, than 
that of A, fluid will pass from A to B by osmosis. But the 
consequence of this passage of water will be to concentrate 
the Na CI in A, so that the partial pressure of this salt in 
A is greater than in B. Na CI will therefore diffuse from 
A to B with the result that the former difference of total 
osmotic pressure will be re-established. Hence there will 
be a continual passage of both water and salt from A to B, 
until B has absorbed the whole of A. This result will 
be only delayed if the osmotic pressure of A is at first 
higher than B, in consequence of a greater concentration 
of Na CI in A. There may be at first a flow of fluid 


from B to A, but as soon as the Na CI concentration on 
the two sides has become the same by diffusion the power 
of x to attract water from the other side will make itself 
felt, and this attraction will be proportional to the osmotic 
pressure of x. 

We have an example of such a process in the absorption 
of salt solutions from the connective tissues by the blood- 
vessels, as well as in the absorption of the normal tissue 
lymph. The capillaries of the connective tissues of the 
limbs and peripheral parts of the body are almost imperme- 
able to proteids. In consequence of this impermeability the 
fluid which is transuded from the capillaries under pressure 
contains very little proteid, whereas it contains exactly the 
same proportion of salts as does the blood plasma. It seems 
probable therefore that the proteid left in solution in the 
capillaries must exert a certain osmotic attraction on the 
salt solution outside the capillaries. It is easy to 
measure this attractive force. If blood serum be placed in 
a small thistle funnel, on the open end of which is stretched 
a layer of membrane soaked in gelatine, and the inverted 
funnel be immersed into salt solution which is isotonic or 
even hypertonic as compared with the serum, measured by 
the freezing-point, within the next two to four days fluid 
will pass into the funnel and rise up in its capillary stem to 
a considerable height. I have found that the osmotic 
pressure of the non-diffusible constituents of blood serum 
measured in this way amounts to between 30 mm. and 40 mm. 
Hg. Now although this osmotic pressure is so small, it is of an 
order of magnitude comparable with that of the hydrostatic 
pressure in the capillaries. This fact is of importance in 
that, whereas the capillary pressure determines transudation 
from the vessels, the effective osmotic pressure of the 
serum (proteids ?) determines absorption by the blood- 
vessels. Moreover the osmotic attraction of the serum for 
the extravascular fluid will be proportional to the force 
expended in the production of this extravascular fluid, so 
that at any given time there must be a balance be- 
tween the hydrostatic pressure in the capillaries and the 
production or absorption of fluid from the extravascular 


spaces — a balance which is known to obtain under physio- 
logical conditions. If we increase the volume of circulating 
fluid we increase intracapillary pressure and the blood 
volume tends to diminish in consequence of increased 
transudation. If we diminish the capillary pressure by 
bleeding the animal, absorption will predominate over exu- 
dation, and the volume of circulating fluid will tend to 
increase towards its normal amount. 

From this cursory study of some of the simplest examples 
of transference of fluids and salts across membranes, we 
may draw certain conclusions as to the main factors which 
are of importance for the process. 

These are : (i) The permeability of the membrane to the 
dissolved substances. This permeability may be of the 
same character as the permeability of water, in which case 
the rates of passage of the dissolved substances across the 
membrane vary as their diffusibilities, and are therefore 
probably some function of their molecular weights. On the 
other hand the membrane may exhibit a certain attraction 
for, or power of dissolving, some dissolved substances to the 
exclusion of others, in which case there will be no relation 
between the diffusibilities and rates of passage of the dis- 
solved substances. 

(2) The osmotic pressure of the solutions. It is evident 
that the rules deduced by Heidenhain from the accepted 
theory of osmotic pressures, and quoted at the beginning of 
this article, are fallacious in consequence of a too narrow con- 
sideration of this second factor to the exclusion of the first. 
At the same time it must be confessed that our knowledge 
of the permeability of different membranes to different 
substances, as well as of the factors on which this per- 
meability depends, is still in an embryonic condition. 
There can be no doubt that a careful exploration of this 
field of research would yield results not only interesting 
to the physicist, but also of incalculable value to the 
physiologist in his investigation of the phenomena of 
living things. 




(i) Heidenhain. Neue Versuche liber die Aufsaugung im Diinn- 
darm. Pfiiigers Archiv, lvi., p. 600, 1894. 

(2) LAZARUS BARLOW. Observations upon the Initial Rates of Os- 

mosis of certain Substances in Water and in Fluids containing 
Albumen. Journ. of Phys., vol. xix., p. 140, 1895. 

(3) ORLOW. Einige Versuche iiber die Resorption in der Bauch- 

hohle. Pfiiigers Archiv, vol. lix., p. 170, 1894. 

(4) Leathes and Starling. On the Absorption of Salt Solutions 

from the Pleural Cavities. Journ. of Phys., vol. xviii., 1895. 

(5) Leathes. Some Experiments on the Exchange of Fluid be- 

tween the Blood and Tissues. Journ. of Phys., vol. xix., p. 1, 


(6) HAMBURGER. Ueber die Regelung der osmotischen Spannkraft 

von Fliissigkeiten in Bauch und Pericardialhohle. Du Bois 
Archiv, p. 281, 1895. 

(7) STARLING. On the Absorption of Isotonic Solutions from the 

Connective Tissues. Journ. of Phys., 1896. 

Ernest H. Starling. 

Science |)ragre$s. 

No. 27. May, 1896. Vol. V. 


IN a discourse to the Members of the Royal Institution 
on the subject of the Metropolitan Water Supply 
nearly thirty years ago, I stated that out of every thousand 
people existing upon this planet at that moment, three 
lived in London ; and, as the population of London 
has in the meantime doubtless grown at a more rapid rate 
than that of the rest of the world, it will probably be no 
exaggeration to say that now, out of every thousand people 
alive on this earth, four live in London ; and therefore any 
matter which immediately concerns the health and comfort 
of this vast mass of humanity may well merit our most 
earnest attention. Amongst such matters that of the 
supply, in sufficient quantity, of palatable and wholesome 
water is certainly not the least in importance. 

It is not therefore surprising that this subject has 
received much attention from several Royal Commissions, — 
notably from the Royal Commission on Water Supply of 
1867, presided over by the Duke of Richmond, the Royal 
Commission on the Pollution of Rivers and Domestic 
Water Supply of Great Britain, presided over by the late 
Sir William Dennison, of which I had the honour to be a 
member ; and lastly the Royal Commission, appointed in 
1892 to inquire into the Water Supply of the Metropolis, 

1 A discourse delivered at the Royal Institution, 21st February, 1896. 



of which Lord Balfour of Burleigh was Chairman, and of 
which Professor Dewar was a member. 

The Royal Institution has also for nearly three-quarters 
of a century been prominently connected with the investiga- 
tion and improvement of the Metropolitan Water Supply ; 
no less than four of our Professors of Chemistry having 
been successively engaged in this work, viz., Professors 
Brande, Odling, Dewar, and myself, whilst three of them 
have been members of the Royal Commissions just 
mentioned. I may therefore perhaps be excused for 
accepting the invitation of our Secretary to bring the 
subject under your notice for the third time. 

On the present occasion I propose to consider it from 
three points of view, viz., the past, the present and the 
future ; and, for reasons which will appear hereafter, I shall 
divide the past from the present at, or about, the year 1883, 
and will not go back farther than the year 1828, when Dr. 
Brande, Professor of Chemistry in the Royal Institution ; 
Mr. Telford, the celebrated engineer ; and Dr. Roget, 
Secretary of the Royal Society were appointed a Royal 
Commission to inquire into the quality and salubrity of the 
water supplied to the Metropolis. 

The Commissioners made careful examinations and 
analyses, and reported as follows : " We are of opinion that 
the present state of the supply of water to the Metropolis 
is susceptible of, and requires, improvement ; that many of 
the complaints respecting the quality of the water are well 
founded, and that it ought to be derived from other sources 
than those now resorted to, and guarded by such restrict- 
tions as shall at all times ensure its cleanliness and purity. 
(At this time the water was pumped from the Thames 
between London Bridge and Battersea.) To obtain an 
effective supply of clear water free from insects and all 
suspended matter, we have taken into consideration various 
plans of filtering the river water through beds of sand and 
other materials ; and considering this, on many accounts, as a 
very important object, we are glad to find that it is perfectly 
possible to filter the whole supply, and this within such 
limits, in point of expense, as that no serious objection can 


be urged against the plan on[ that score ; and with such 
rapidity as not to interfere with the regularity of service." 

Before the year i82g, therefore, the river water supplied 
to London was not filtered at all ; but after the issue of 
this report, the Companies set themselves earnestly to work 
to improve the quality of the water by filtration. 

The first filter, on a working scale, was constructed and 
brought into use by the Chelsea Water Company in the 
year 1829. But even as late as 1850 only three out of the 
seven principal companies filtered the river water which 
they delivered in London ; and it was not until 1856 that 
filtration was made compulsory by Act of Parliament, 
whilst it can scarcely be doubted that, between this date and 
the year 1868, when my observations on turbidity were first 
commenced, the operation was very imperfectly performed. 

In the year 1832, and again in 1849, London was 
severely visited by epidemic cholera, and the agency of 
drinking water in spreading the disease forced itself upon 
the attention of the observant portion of the medical pro- 
fession. It was Dr. Snowe, however, who in August, 
1849, first formally enunciated the doctrine that drinking 
water polluted by choleraic matters is the chief mode by 
which cholera is propagated. 

Received at first with incredulity, this doctrine was 
supported by numerous facts, and it soon caused renewed 
attention to be directed to the quality of the water then 
being supplied to the Metropolis ; with the result that the 
intakes of the various Companies drawing from rivers 
were, one after another, removed to positions above the 
reach of tidal influence ; the Thames water being with- 
drawn from the river above Teddington Lock, and the Lea 
water at Ponder's End, above the tidal reaches of that river. 

In every visitation of Asiatic cholera to London, the 
water supply was either altogether unfiltered or imperfectly 
filtered, besides being derived from highly polluted parts 
of the Thames and Lea ; and the enormous loss of life, 
amounting in the aggregate to nearly 36,000 people, can 
only be attributed to this cause. It has been abundantly 
proved that efficient filtration is a perfect safeguard against 


the propagation of the disease, and since the year 1854 no 
case of Asiatic cholera in London has been traced to the 
use of filtered river water. 

These are the results arrived at by the most general 
investigation of the subject. They show that in every 
epidemic, the mortality varied directly with the intensity of 
the drainage pollution of the water drunk by the people ; 
but if time permitted, a more detailed study of the 
statistics in both epidemics would demonstrate, much more 
conclusively, this connection between cholera mortality and 
the pollution of drinking water — a connection which has quite 
recently been terribly emphasised in the case of Hamburg. 

Such is the verdict with regard to cholera, and the 
same is true of that other great water-borne disease typhoid 
fever. But, unlike cholera, this disease is disseminated in 
several other ways, and its presence or absence in any 
locality may not, of necessity, have any connection with 
drinking water, as is strikingly shown by the health 
statistics of Manchester. 

There is no evidence whatever that, since the year 
1869, when typhoid fever appeared for the first time as a 
separate disease in the Registrar General's reports, it has 
been conveyed by the water supply of the Metropolis. 
An inspection of the diagram (No. 1) shows, it is true, a 
greater proportional mortality during the period of imperfect 
filtration than during the later period ; that is to say from 
1883 when the process began to be performed with uniform 
efficiency ; but the plotting of a similar curve for the deaths 
by typhoid in Manchester shows that this disease arises from 
other causes than polluted water, since the water supply 
of Manchester, derived as it is from mountain sources, is 
above all suspicion of this kind. These other causes have 
during the last ten years been much mitigated in London 
by various sanitary improvements ; whilst, as shown in 
the diagram, there has been no corresponding mitigation 
in Manchester. 

Although very soon after the year 1856 all the water 
supplied to the Metropolis was obtained from sources much 
less exposed to drainage pollution, it was still very carelessly 



filtered. Previous to the year 1868, there are no records 
of the efficiency, or otherwise, of the filtration of the 
Metropolitan water supply derived from rivers, as dis- 
tinguished from deep wells, the water of which is perfectly 
clear without filtration. 

It was in the year 1868 that I first began to examine 
the water supplied to the Metropolis from rivers with 
reference to efficiency of filtration. I n that year, out of 
eighty-four samples examined, seven were very turbid, 
eight turbid, and ten slightly turbid, so that altogether no 
less than nearly 30 per cent, of the samples were those of 
inefficiently filtered water. The Metropolitan Water Supply 
then, up to the year 1868, may be shortly described as 

1 rues fc ifigft am MUtCCCTCI 

. contrasted w.n Tusrt'.rr 

1 1 

4 4-4- -- 



-A—_ A 

T A 

\— A ti 

Si J. 

\ / v ^" 

44 si 

\ l 4h 

V f 


J_ v ' - - -J ^. 

No. 1. 

derived for many years from very impure sources with 
either no filtration at all, or with very inefficient filtration ; 
and afterwards, when the very impure sources were 
abandoned, the supply was still often delivered in a very 
inefficiently filtered condition. But, after the establishment of 
monthly reports on the filtration of the river-derived supplies, 
the quality of these waters gradually improved in this most 
important respect, as is seen from diagram No. 1. In this 
diagram, the continuous line with dots represents the 
mortality from typhoid in Manchester, the broken and 
eroped line the contemporaneous mortality in London, and 
the dotted curve the degree of turbidity of the London 
water supply. 


These observations graphically represented in the dia- 
gram show that, at the time they were commenced, the 
filtering operations were carried on with considerable care- 
lessness, and that this continued, though to a less extent, 
down to the year 1883, since which time, and especially 
since 1884, the efficiency of filtration of all the river waters 
supplied to the Metropolis has left little to be desired. 

What is it then that separates the past from the present 
water supply of London ? In the first place there is the 
change of source — I mean the change in position of the 
intakes of the several Companies drawing from the Thames 
and Lea — and the total abandonment of the much-polluted 
river Ravensbourne by the Kent Water Company. So 
long as the water supply was derived from the tidal reaches 
of the Thames and Lea, receiving as these reaches did the 
drainage of immense populations, the risk of infection from 
water-borne pathogenic organisms could scarcely be other- 
wise than imminent ; for, although we now know efficient 
filtration to be a perfect safeguard, anything short of effi- 
ciency must be attended with risk in the presence of such 
extreme pollution. 

Nevertheless, the line of demarcation between the past 
and the present water supply of the Metropolis is, in my 
opinion, to be drawn, not when the intakes of the river 
companies were removed to positions beyond the possibility 
of pollution by the drainage of London, but it must be drawn 
at the time when efficient filtration was finally secured and 
ever since maintained, that is to say, in the year 1884. 

The removal of turbidity by sand filtration, however, 
refers only to suspended matters ; but there are sometimes 
objectionable substances in solution of which organic matter 
is the most important. River water and mountain water, 
even when efficiently filtrated, contains more organic matter 
than spring or deep well water; but this is reduced in quantity 
by storage and especially by filtration, although these waters 
can perhaps never be brought up to the standard of organic 
purity of spring and deep well water. 




At present London is supplied with water from four 
sources — the Thames, the Lea, the New River, and deep 
wells. Of these the deep wells yield as a rule the purest 
water, requiring no filtration or treatment of any kind before 
delivery for domestic use. The river waters, on the other 
hand, require some kind of treatment before delivery — 
storage, subsidence in reservoirs, and filtration. The water 
from the Thames is abstracted at and beyond Hampton, far 
above the reach of the tide and London drainage. The 
water from the Lea is taken out at two points, viz., at Angel 
Road near Chingford, by the East London Water Company, 


RT 10 




1 W 










■ O 










-r — 











k : v ; 


■■ ■ 











No. 2. 

and above Hertford by the New River Company, who 
convey it to Green Lanes by an open conduit twenty-five 
miles long, called the New River Cut, in which it is mixed 
with a considerable volume of spring and deep well water. 

All three river waters are affected by floods and are, as 
raw materials, of considerably different quality as regards 
organic purity (see diagram No. 2). From these raw 
materials by far the largest volume of the Metropolitan 
Water Supply is derived, and the chemical or organic 
purity of the water sent out to consumers stands in direct 
relation to the organic purity of the raw material used, as 



is seen from the diagrams Nos. 3, 4 and 5, which show the 
proportional amounts of organic elements in the raw and 
filtered waters ; they also show the advantage of storage 
in excluding flood water, No. 4 shows that floods in March 









t* 1 



i 1 

























-■ ■» ' * . XT 

No. 3. 

arid August were circumvented, but not in November. The 
numbers in the margins of the diagrams express the pro- 
portional amount of organic elements, that in the Kent 


























\ / 









DN ■' 




No. 4. 

Company's water during the nine years ending December 

1876 being taken as unity, as is depicted in diagram No. 5. 

Hitherto I have spoken of chemical purity or comparative 

freedom from organic matter only, but the spread of diseases 



such as cholera and typhoid fever through the agency of 
drinking water has no connection whatever with the chemical 
or organic purity of the water. These diseases are propa- 
gated by living organisms of extreme minuteness, to which 
the names bacilli, bacteria, and microbes have been given, 
and here comes the important question how, if at all, does 
filtration secure immunity from these water-borne diseases ? 
To Dr. Koch of Berlin, we are indebted for the answer 
to this question. By his discovery of a means of isolating 
and counting the number of bacteria, or bacilli, or microbes 
and their spores in a given volume of water, we were, for 
the first time, put into possession of a method by which the 
condition of water as regards these living organisms, before 




No. 5. 

and after filtration, can be determined with quantitative 
exactness. The enormous importance of this invention 
(which was first made known and practised in England in 
1882 by the late Dr. Angus Smith) is evident, when it is 
borne in mind that the living organisms, harmful or harm- 
less, contained in water are of such extreme minuteness as 
practically to defy detection by ordinary microscopical 
examination. But although the microscope cannot detect 
with certainty single bacteria or their spores, even the 
naked eye can easily discern towns or colonies consisting of 
thousands or even millions of such inhabitants. 

Dr. Koch's method accomplishes at once two things : 
it isolates, in the first place, each individual microbe or 


germ ; and, secondly, places it in conditions favourable for 
its multiplication which takes place with such amazing 
rapidity that, even in a few hours, or at most in two or 
three days, each organism will have created around itself 
a visible colony of innumerable members — a town in fact 
comparable to London itself for population. 

By operating upon a known volume of water, such as 
a cubic centimetre for instance, the number of separate 
organisms or their spores, in a given volume of the water 
under investigation can thus be determined. The following 
is the method now adopted in carrying out Koch's process 
for the bacterial investigation of drinking water : — 

i. Preparation of the nutritive medium. 

2. Sterilisation of the medium. 

3. Collection of the sample of water in a vacuous tube 

afterwards to be hermetically sealed. 

4. Transport of the sample to the bacteriological 

laboratory, packed in ice to prevent multiplica- 

5. Mixture of a known volume of the water sample 

with the nutrient medium. 

6. Casting of the mixture into a solid plate. 

7. Incubation of the solid plate. 

8. Counting of the colonies. 

9. Examination of separate colonies, or rather of the 

individual members under the microscope. 

Sometimes the cultivations are made upon a plate of 
the substance called agar which resembles isinglass, and 
bears a temperature of blood heat without melting. 

In order to ascertain the effect of filtration upon the 
bacterial quality of water, it is absolutely necessary that 
the sample should be taken immediately after it has passed 
through the filters ; for, if it be obtained from the delivery 
mains in town, that is to say, after the water has passed 
through many miles of pipes, the rapid multiplication of 
these organisms, except in very cold weather is such, that 
a water which contains only a single living organism per 
cubic centimetre, as it issues from the filter, may contain 



100 or 1000 in the same volume when, after several hours, 
it arrives on the consumer's premises. 

Now what is the effect of sand filtration as carried out 
by the various Water Companies supplying London upon 
the living matter contained in the raw river water ? // is 
simply astounding : water containing thousands of bacteria 
per cubic centimetre, for a single drop of Thames water 
sometimes contains nearly 3000 separate living organisms, 
comes out from the sand filters with fifty, thirty, ten, or even 
less of these organisms per cubic centimetre, or the number 
of microbes in a single drop is reduced to two or even to zero. 

WATER 1894. 







No. 6. 


Rather less than one-tenth of the total volume of water 
supplied to London is derived by the Kent Water Company 
from deep wells in the chalk. As it issues from the porous 
rock into the fissures and headings of these wells, this water 
is, in all probability, absolutely sterile ; but by the time it 
has been pumped up to the surface it usually contains a 
certain number, though small, of microbes. Thus, during 
the year 1892 it contained on the average six per cubic 
centimetre in 1893, thirteen; in 1894, fifteen; and in 
1895, eight. 

The diagram No. 6 shows graphically the bacterial 
improvement of the Thames water by filtration during 


the year 1894. ^ n this diagram the black squares represent 
the number of microbes in a given volume of the raw water 
in each month, and the white centres the number remaining 
in the same volume after filtration. 

Although deep well water has, from a bacterial point 
of view, a decided advantage, the filtered river waters 
are not very far behind, and there is every reason to believe 
that with the improvements which are now being carried out 
by the various river Water Companies, the Kent Company's 
deep well water will, before long, be run very hard by the 
other supplies. 

By the examination of the water as it issues from the 
filters, the utmost freedom from microbes, or maximum 
degree of sterility of each sample is determined. This 
utmost freedom from bacterial life after all sources of con- 
tamination have been passed is obviously the most 
important moment in the history of the water ; for the 
smaller the number of microbes found in a given volume at 
that moment the less is the probability of pathogenic or 
harmful organisms being present ; and although the non- 
pathogenic may afterwards multiply indefinitely this is of 
no consequence in the primary absence of the pathogenic ; 
but it is only fair, in describing the character of the present 
water supply of London, to say that not a single pathogenic 
organism has ever been discovered even in the ^filtered 
water as it enters the intakes of the various Companies, 
although these organisms have been carefully sought for. 
It is sometimes said that the non-pathogenic organisms 
found in water may be beneficial to man ; but this idea is 
not borne out by the fact of their entire absence from the 
food which nature provides for young animals. Healthy 
milk is absolutely sterile. 

As it is at present impracticable to obtain water, uni- 
formly at least, free from microbes, it is desirable to adopt 
some standard of bacterial purity ; and 100 microbes per 
cubic centimetre has been fixed upon by Dr. Koch and 
myself as the maximum number allowable in potable water. 
This standard is very rarely infringed by the London 
Water Companies, whilst I have every reason to hope that, 


in the near future, now that special attention is directed to 
bacterial filtration, it will not be approached within 50 per 
cent. This hope is based not only upon my own observations, 
but also upon the exhaustive and exceedingly important 
investigations carried out at the Lawrence Experiment 
Station by the State Board of Health of Massachussetts, 
under the direction of Mr. George W. Fuller, the official 
biologist to the Board. 

More than six years have already been spent in the prose- 
cution of these American experiments, and many thousands 
of samples of water have been submitted to bacterial cultiva- 
tion. The Massachussetts experimental filters are worked at 
rates up to 3,000,000 gallons per acre daily, which renders 
the results available for application to public water supplies ; 
indeed none of the water delivered in London is filtered at 
so rapid a rate as this. It was found that at these rates all 
the disease-producing germs, which were intentionally and 
in large numbers added to the unfiltered water, were 
substantially removed. The filters were so constructed 
and arranged as to allow direct comparison of the bacterial 
purification of water under different rates of filtration, with 
sand of different degrees of fineness, with different depths 
of the same sand, and with intermittent and continuous 

The actual efficiency of these filters was also tested by 
the application of the bacillus of typhoid fever. Very large 
numbers of these bacilli and of other species were applied 
in single doses to the several filters at different times, and 
the effluent was examined four times daily for several days 
afterwards. The results so obtained give a thoroughly 
trustworthy test of the degree of bacterial purification 
effected by each of the experimental filters, and these are 
the data which have been largely used by the Mas- 
sachussetts State Board of Health in deducing the rules 
which they consider ought to be observed in water filtration. 

Among the subjects investigated by means of these 
experimental filters were : — 

1. The effect, upon bacterial purification, of the rate of 


2. The effect of size of sand grains upon bacterial 


3. The effect of depth of material upon bacterial 


4. The effect of scraping the filters upon bacterial 

Time does not permit of my giving the answers to these 
questions in detail ; but they may be summarised as follows : — 

1. The rate of filtration between 500,000 and 3,000,000 
gallons per acre per day exercises practically no effect on 
the bacterial purity of the filtered water. It is worthy of 
note that the rates of filtration practised by the several 
Water Companies drawing their supplies from the Thames 
and Lea are as follows: Chelsea Company, 1,830,000; 
West Middlesex, 1,359,072 ; Southwark Company, 1,568, 160; 
Grand Junction Company, 1,986,336; Lambeth Company, 
1,477,688; New River Company, 1,881,792; and East 
London Company, 1,393,920. Hence not one of the 
London Companies filters at the rate of 2,000,000 gallons 
per acre per day ; at which rate in the Massachussett's 
filters 99*9 per cent, of the microbes present in the raw 
water were removed. 

2. The effect of size of sand grains was found to be very 
considerable ; and, in confirmation, I find that by the use of 
a finer sand than that employed by the Chelsea Company, 
the West Middlesex Company is able, with much less stor- 
age, to attain an equal degree of bacterial efficiency. 

3. The depth of sand, between the limits of one and five 
feet, exercises no practical effect on bacterial purity when 
the rate of filtration is kept within the limits just specified. 
And this result is quite borne out by my own experience 
gained in the bacterioscopic examination of the filtered 
waters of the seven Companies supplying the Metropolis 
from rivers. Thus the New River Company, with i*8 feet 
of sand on the filters, compares favourably with the Chelsea 
Company, the sand on whose filters is more than twice that 

Placed in the order of thickness of sand on their filters, 
the Metropolitan Companies range as follows : Chelsea, 


Lambeth, West Middlesex, Southwark, East London, 
Grand Junction, and New River. Placed in the order of 
efficient filtration they range as follows : Chelsea and West 
Middlesex equal, New River, Lambeth, East London, 
Southwark, and Grand Junction. 

4. When there is such an accumulation of deposit on 
the surface of a sand filter that, for practical purposes, 
sufficient water cannot be made to pass through it, the 
surface of the filter has to be scraped ; that is to say, the 
mud and about half an inch of the sand are removed from 
the surface. After this operation, there is sometimes an 
increase in the number of bacteria in the filtered water, and 
it was noticed that the increase was greater in shallow than 
in deep filters and with high than with low rates of filtration; 
and there is no doubt that the effect of scraping is con- 
siderably magnified when coarser descriptions of sand are 
employed, as in the case of the filters of the London Water 
Companies. I should like, therefore, to impress upon the 
engineers of these Companies the desirability of using finer 
sands than are at present employed. 


I have found that the number of bacteria in a given 
volume of filtered water is to a considerable extent in- 
fluenced by the number contained in the raw water supplying 
the filter ; and from this point of view, therefore, the bacterial 
condition of the raw river water used in the Metropolis is of 
no inconsiderable importance. 

Since May, 1892, I have made monthly determina- 
tions of the number of microbes capable of developing 
on a gelatine plate in a given volume of raw Thames water 
collected at the intakes of the Metropolitan Water Companies 
at Hampton ; and the number has varied during this time 
between 631 and 56,630 per cubic centimetre, the highest 
numbers having, as a rule, been found in winter or when 



the temperature of the water was low, and the lowest 
in summer or when the temperature was high. 

Now, besides temperature, there are two other conditions 
to either of which this difference may be attributed, viz. y 
sunshine and rainfall, and I have endeavoured by a series 
of graphic representations to disentangle these possible 
influences from each other by placing the results of the 
microbe determinations in juxtaposition with (i) the tem- 
perature of the water at the time the samples were taken ; 

^ 3b 

g V 



Jly Mb Sep 

Oct Hot Ok Jin Ftk Mil 


Atd Ha Jun.Jf 




And Sep Oct Nov Dei 

1 * 







* ■*• 



i oooo 

IB 000 

a. ooo 











, — 



u— 1 






r" ' 

- -\ 




i _L 







T 1 











, ' 





- 1 







\ t 








, j__ 


























No. 7. 

(2) the number of hours of sunshine on the day and up to 
the hour when each sample was drawn and on the two 
preceding days, and (3) the flow of the Thames over 
Teddington Weir on the same day expressed in millions 
of gallons per twenty-four hours. Although the graphic 
representations were confined to the Thames, the conditions 
affecting bacterial life in this river are doubtless equally 
potent in other rivers and streams. 

The samples for microbe cultivation were collected at 



about nine inches below the surface of the water in partially 
exhausted and sealed tubes, the ends of which, when the 
tubes were lowered to the required depth, were broken off 
by an ingenious contrivance devised by my Assistant, Mr. 
Burgess. On being withdrawn from the river the tubes 
were immediately hermetically sealed and packed in ice for 
conveyance to my laboratory, where the cultivation was 
always commenced within four hours of the time of collection. 


dun JuJi 

Auj Sep Oct /lot DecJin 


Feb. MrMprl % dun My A ub Sep Oct Hot Dec Jail 


"« *rt 4»r/% i/w iW« A//> Sep. Oct Nov Dec 































1 * 
1 1 

! i 




1 j 







1 1 
' t 











r ft: j 








\ / 



L < 

. 1 





\ 1 







l! /! | 


1 1 

'^t ; 







Tl l 





III ' 





i 1' 

»! : 




■»,' " 









1 1 

/ 1 

/ 1 






i' 1 

; ;i 




\\\ ; 




"1 1 


1 \\i 1 





] ; 













/ 1 ^ 


\ 1 



1 ! 


. 1 . 

No. 8. 

For the records of sunshine, I am indebted to the 
kindness of Mr. James B. Jordan of Staines ; and for 
gaugings of the Thames at Teddington Weir to Mr. C. J. 
More, the engineer to the Thames Conservancy Board. 

The graphic representation of these collateral observa- 
tions affords definite evidence as to which of the three 
conditions — temperature, sunshine, and Mow of the river — 
has the predominant influence upon bacterial life in the 
water. The first diagram (No. 7) compares the number of 




microbes per cubic centimetre with the temperature at the 
time the sample was taken. The horizontal lines express 
the numbers of microbes and the temperature, while the 
vertical lines denote the months when the samples were 
taken. For obvious reasons the horizontal lines express- 
ing the numbers of microbes and temperatures are num- 
bered in opposite directions. 

The diagram shows that although coincidences between 
a high number of microbes and a low temperature are not 



Ms*. Jin Jlr. M 5a Oct. Hot 

Occihn CeAMir. 






Oct NwDutkn febthr Apit 'MfyJun. 




9 00 

a po 


V) Hi 




| " 



1 1 

1 1 







1 1 






i \ 





















! » 


' \ 








































! ' 


No. 9. 

wanting, some other condition entirely masks the effect, if 
any, of temperature. 

The next diagram (No. 8) institutes the comparison 
between the number of microbes and the hours of sun- 
shine to which the water has been exposed. The diagram 
is constructed on the same lines as the first. 

It is here seen that, as in the case of temperature, 
there is some other condition which entirely overbears the 
influence of sunlight in the destruction of microbes in the 


river water. This condition is the amount of rainfall higher 
up the river, or, in other words, the volume of water flowing 
along the river bed, as is seen from the comparison repre- 
sented in the next diagram (No. 9). 

This diagram shows very conclusively that the volume 
of water flowing in the Thames is the paramount influence 
determining the number of microbes. It compares the 
volume of water in the river gauged at Teddington Weir 
with the number of microbes found in the raw Thames 
water at Hampton on the same day. In this diagram, the 
numbers representing the flow of the river in millions of 
gallons per day and the number of microbes per cubic 
centimetre in the water both run from the bottom of the 
diagram upwards. 

Comparing the curves in the diagram it is seen that, 
with very few exceptions, a remarkably close relation is 
maintained between them. 

The only exception of any importance to the rule that 
the number of microbes varies directly with the flow of 
the river, occurring during the thirty-two months through 
which these observations were continued, happened in 
November, 1892, when the flow increased from 501 mil- 
lions of gallons in October to 1845 millions in November, 
whilst the microbes actually diminished in number from 
2216 to 1868 per cubic centimetre. Neither the sunshine 
nor the temperature records of these two months, however, 
afford any explanation of this anomalous result, for there 
was a good deal of sunshine in October before the collection 
of the sample and the temperature was higher, whilst in 
November no ray of sunshine reached the Thames during 
the three days preceding the taking of the sample and the 
temperature was nearly 4 C. lower than in the preceding 
month. I have ascertained, however, that the Thames 
basin had been twice very thoroughly washed out by heavy 
floods before the time when the November sample was 
taken, and this affords a satisfactory explanation of the 
anomalous result yielded by this sample. 

These comparisons demonstrate that the number of 
microbes in Thames water depends directly upon the rate 



of flow of the river, or, in other words, on the rainfall, and 
but slightly, if at all, upon either the presence or absence of 
sunshine or a high or low temperature ; and they are con- 
firmed by the continuation of these observations during the 
year 1895 exhibits in diagram No. 10. 

With regard to the effect of sunshine upon bacterial 
life, the interesting observations of Dr. Marshall Ward 
leave no doubt that sunlight is a powerful germicide ; still 
it is obvious that its potency in this respect must be greatly 
diminished, if not entirely annulled, when the solar rays 
have passed through a stratum of water of even com- 



^ < 

O u 




34 00c 









































1 ; 







11 5 


















2 COO 




.._ i 






No. 10. 

paratively small thickness before they reach the living 
organisms. By a series of ingeniously contrived experi- 
ments, Mr. Burgess has demonstrated the correctness of 
this view. 

A sterile bottle about half filled with Thames water was 
violently agitated for five minutes to insure equal distribu- 
tion of the organisms. Immediately afterwards a number 
of sterile glass tubes were partially filled with this water 
and sealed hermetically. Three of these tubes were 
immediately packed in ice, and the remainder were attached 
in duplicate at definite distances apart to a light wire frame 
which was then suspended vertically in the river. The 


experiments were made near the Grand Junction Company's 
Intake at a place favourable for the sun's rays to fall on the 
river without any obstruction. 

The river was at the time in a very clear condition and 
contained but little suspended matter ; whilst the day was 
fine, although clouds obscured the sun occasionally. The 
tubes were exposed to light in the river for four and a half 
hours — from ICV30 a.m. to 3 r.M. on 15th May, 1895. At 
the end of this time the tubes were packed in ice for trans- 
port to my laboratory, where the cultivation was started 
immediately. The colonies were counted on the fourth day 
and yielded the results given in the table : — 

No. of Colonies 
per c.c. 
Thames water packed in ice immediately after collection - 2127 
Thames water after exposure to sunlight for 4^ hours at 

surface of river ------- 1140 

Thames water after exposure to sunlight for 4^ hours at 

6 in. below surface of river ----- i^o 

Thames water after exposure to sunlight for 4^ hours at 

1 ft. below surface of river - - - - - 2150 
Thames water after exposure to sunlight for 4J hours at 

2 ft. below surface of river ----- 2430 
Thames water after exposure to sunlight for 4! hours at 

3 ft. below surface of river ----- 2440 

These experiments show that, on 15th May the 
germicidal effect of sunlight on Thames microbes was nil 
at depths of one foot and upwards from the surface of the 
water. It cannot, therefore, excite surprise, that the effect 
of sunshine upon bacterial life in the great mass of Thames 
water should be nearly, if not quite, imperceptible. It is 
thus ascertained that sunlight can only kill the germs, or 
microbes, near the surface of the water, whilst those at any 
depth, for the most part, escape destruction. 

On the other hand the enormous effect of floods in 
augmenting the number of microbes can hardly surprise us, 
for when a great body of water has flowed over the banks 
of the river, which are at other times dry and exposed, it 
carries along with it countless impurities — an effect common 
both to the main stream and its tributaries. The Thames 


basin is as it were, on every such occasion, thoroughly- 
washed out, and it is only to be expected that the number 
of microbes in the water should be enormously increased as 
is found to be the case. 
Now with respect to 


In view of the rapid increase of the population of Lon- 
don, fears have from time to time been entertained that the 
water supply from the Thames basin, that is to say from 
the rivers Thames and Lea, supplemented by water from 
springs and deep wells within the basin itself, would soon 
be insufficient in quantity ; whilst the quality of the 
water taken from the river has, up to a comparatively 
recent date, been considered unsatisfactory. On these 
grounds various schemes have, from time to time, been 
brought forward for the supply of the Metropolis from 
other river basins — from the Wye, the Severn, the river 
basins of North Wales, and of the Lake Districts of 
Cumberland and Westmoreland. 

It is worthy of note, however, that all the Royal Com- 
missions have arrived tinanimously at the conclusion 
that the quantity of water obtainable from the Thames 
basin is so ample as to render the necessity of going else- 
where a very remote contingency. 

I shall now endeavour to put, very shortly before you 
a few facts which, in my opinion, prove that, both as 
regards quantity and quality, the Thames basin will for a very 
long time to come afford an abundant supply for the Metro- 
polis. There is indeed no river basin in Great Britain 
which affords such an abundant supply of excellent water 
as that available in the Thames basin. 

Besides that which flows directly into the rivers, this 
water is contained in the Chalk, Oolite, and Lower Green- 
sand, which are the best water bearing strata in the 
kingdom. From these rocks it issues in copious springs 
of unsurpassed organic purity. I have personally inspected 
every spring of importance in the Thames basin and have 



analysed samples of the water. The results, in a very con- 
densed form, are recorded in the annexed Table. Twenty- 



Results of Analysis in 
Parts per 100,000. 


Average of 21 




Average of 5 




Average of 8 



Average of 36 


Total Saline Matters - 
Organic Carbon - - 
Organic Nitrogen - - 
Hardness before boiling 





5 - 5 




2 5'3 






one samples of Oolitic spring - water were analysed, and 
every one of these was of even greater organic purity 
than the water delivered by the Kent Company, which I 
have always regarded as the standard of organic purity to be 
aimed at in all other Water Works. 

Five springs issuing from the Lower Greensand were 
examined ; and again, every one of these was of even 
greater purity, organically, than the Kent Company's 
water ; whilst they were, on the average, only one-third as 
hard. Forty-six samples of water from the Chalk were 
chemically examined, and these also contained but the 
merest traces of organic matter. 

All these samples from the Chalk were derived from 
sources where the water-bearing- stratum is free from a 
covering of London clay ; but, as soon as the Chalk dips 
beneath the London Tertiary Sands and clay, the quality of 
the water undergoes a remarkable alteration. The total 
solids in solution are greatly increased in amount, whilst 
the hardness is much mitigated, owing to the replacement 
of bicarbonate of lime by bicarbonate of soda. These 
waters are also of high organic purity ; but, as the quantity 
is very limited, it is useless to dwell upon them. They 


supply the Trafalgar Square fountains and the London 
breweries, and we can well afford to leave them to be con- 
verted into beer. For dietetic purposes there is no better 
water in the kingdom than the underground water of the 
Thames basin. For sentimental reasons I should like to 
see it conveyed to the works of the various companies in 
special conduits ; but we have seen that, on hygienic 
grounds, it may safely be allowed to flow down the bed of 
the Thames if it be afterwards efficiently filtered. 

So much for quality, now as to quantity ; the basins of 
the Thames and Lea include an area of upwards of 5000 
square miles. Of this rather more than one half (including 
the Oolitic, Cretaceous, and portions of the Tertiary Forma- 
tions) is covered by a porous soil upon a permeable water 
bearing stratum. The remainder is occupied by the 
Oxford, Kimmeridge, Gault, and London Clays ; being 
thus covered by a clay soil upon a stiff and impervious 

The annual rainfall of the district is estimated at an 
average of twenty-eight inches. The rivulets and streams 
of the Thames basin are formed and pursue their course on 
clay land. There are no streams on the Chalk. That 
which falls upon this porous stratum and does not evaporate 
sinks, mostly where it alights, and heaps itself up in the 
water-bearing stratum below, until the latter can hold no 
more. The water then escapes as springs at the lowest 
available points. 

Innumerable examples of these springs occur all round 
the edge of the Thames basin, and at various points within 
it. Thus from the Chalk they are ejected at the lip of the 
Gault ; and in the Oolitic area by the Fuller's Earth below 
it, or by the Oxford Clay, geologically, above it. 

According to the guagings of the engineer of the 
Thames Conservancy Board there passed over Teddington 
Weir, in 1892, 387,000 millions of gallons, equal to an 
average flow of 1060 millions of gallons daily. In the 
following year, 1893, their passed over Teddington Weir 
an aggregate of 324,227 millions of gallons, or a daily 
average of 888 millions, the average for the two years being 


974 millions of gallons, and this number does not in- 
clude the 1 20 millions daily abstracted by the five London 
Water Companies who draw their supplies from the 

Thus, in round numbers, we may say that after the 
present wants of London have been supplied from this 
river, there is a daily average of nearly 1000 millions of 
gallons to spare. Surely it is not too violent an assumption 
to make that the enterprising engineers of this country can 
find the means of abstracting and storing for the necessary 
time one-fourth of this volume. 

As regards the quality of this stored water, all my 
examinations, of the effect of storage upon the chemical and 
especially upon the bacterial quality, point to the conclusion 
that it would be excellent. Indeed the bacterial improve- 
ment of river water by storage for even a few days is 
beyond all expectation. Thus the storage of Thames water 
by the Chelsea Company for only thirteen days reduces 
the number of microbes to one-fifth the original amount, 
and the storage of the river Lea water for fifteen days, 
by the East London Company, reduces the number on the 
average from 9240 to i860 per cubic centimetre or to one- 
fifth ; and lastly, the water of the New River Cut, con- 
taining on the average 4270 microbes per cubic centimetre 
contains, after storage for less than five days, only 18 10, 
the reduction here being not so great, partly on account 
of the shorter storage, but chiefly because the New River 
Cut above the point at which the samples were taken, is 
itself a storage reservoir containing many days' supply after 
filtration. Indeed quietness in a subsidence reservoir is, 
very curiously, far more fatal to bacterial life than the most 
violent agitation in contact with atmospheric air ; for the 
microbes which are sent into the river above the falls of 
Niagara, by the City of Buffalo, seem to take little or no 
harm from that tremendous leap and turmoil of waters, 
whilst they subsequently, very soon, almost entirely dis- 
appear in Lake Ontario. 

It is not, therefore, too much to expect that storage for, 
say a couple of months, would reduce the number of 


microbes in the Thames flood water down to nearly the 
minimum ever found in that river in dry weather, whilst, 
by avoiding the first rush of each flood, a good chemical 
quality could also be secured. 

There is, therefore, I think, a fair prospect that the 
quantity of water derivable from the Thames at Hampton 
could be increased from its present amount (120 millions of 
gallons per diem) to 370 millions. 

Again, in the river Lea, although here the necessary 
data for exact calculations are wanting, it may be assumed 
that the present supply of 54 millions of gallons could 
be increased by the storage of flood water to 100 
millions per day. To these volumes must be added the 
amount of deep-well water which is attainable from those 
parts of the Thames basin which lie below Teddington Lock, 
and in the Lea basin beloiv Lea Bridge, and which was 
estimated by the last Royal Commission at rather more 
than 67J millions of gallons. 

Thus we get the grand total of 53735- millions of 
gallons of excellent water obtainable within the Thames 
basin, the quality of which can be gradually improved, if it 
be considered necessary, by pumping from the water bear- 
ing strata above Teddington and Lea Bridge respectively, 
instead of taking the total supply from the open rivers 
above these points. Such a volume of water would scarcely 
be required for the supply of the whole water area of Lon- 
don at the end of fifty years from the present time, even 
supposing the population to go on increasing at the same 
rate as it did in the decade 1881-91, which is an assumption 
scarcely likely to be verified. 

In conclusion, I have shown that the Thames basin can 
furnish an ample supply for fifty or more years to come, 
whilst the quality of the spring and deep-well waters and of 
the filtered river water would be unimpeachable. To secure 
these benefits for the future, storage must be gradually pro- 
vided for 1 1,500 millions of gallons of flood water judiciously 
selected in the Thames Valley, and a proportionate volume 
in the basin of the Lea ; whilst filtration must be carried to 
its utmost perfection by the use of finer sand than is at 


present employed, and by the maintenance of a uniform rate 
during the twenty-four hours. 

There is nothing heroic in laying pipes along the banks 
of the Thames, or even making reservoirs in the Thames 
basin. They do not appeal to the imagination like that 
colossal work, the bringing of water to Birmingham from 
the mountains of Wales, and there is little in such a scheme 
to recommend it to the minds of the enterprising engineers 
of to-day. Nevertheless, by means of storage, by utilising 
springs, by sinking deep wells, and by such comparatively 
simple means, we have, in my opinion, every reason to con- 
gratulate ourselves that for half a century at least we have 
at our dooi's, so to speak, an ample supply of water which for 
palatability, wholesomeness, and general excellence will not 
be surpassed by any supply in the world. 

E. Frankland. 


THE literature relating to this group of worms is 
summed up in my Monograph of the Oligochceta 
lately issued by the Clarendon Press ; but so energetic are 
the unfortunately somewhat few workers in this particular 
subject that new facts have gone on accumulating with some 
rapidity since the publication of that work. It is my 
intention in the present article to offer the reader a re'sume 
of this latest work with, naturally, some references to what 
has gone before. 

It is agreed by all those who are acquainted with the 
terrestrial Oligochaeta that their peculiar mode of life, their 
susceptibility to sea water, and the comparatively few 
chances of dispersal enjoyed by them, render their distribu- 
tion highly important in estimating the relations between 
land masses now and in the past. This has an especial 
bearing upon the theory of the former northward extension 
of the Antarctic Continent, a matter upon which much has 
been written lately. To deal adequately with this large 
question would of course demand more space than can be 
allowed me. I shall content myself with referring solely to 
the evidence which is forthcoming from the study of earth- 
worms. Fortunately we are in possession of a considerable 
amount of information about the terrestrial Olioochaeta of 
New Zealand and Patagonia ; the former country indeed 
must be regarded as being better known perhaps than any 
quarter of the globe, excepting of course Europe. The 
extensive collections lately made by Dr. Michaelsen in 
South America have added largely to the number of species 
brought back by his predecessors. It results from an 
examination of the species found in the two countries that 
in both of them the prevailing types belong to the genera 
Acanthodrilus and Microscolex, particularly the former. Of 
the thirty-two indigenous species at present known from 
Patagonia and the more southern parts of the South Ameri- 


can Continent, twenty are members of the genus Acantho- 
drilus, eleven are Microscolex and one is a PericJiesta. Besides 
these are a few obviously imported Lumbricus and 
Allolobophora from Europe or North America. I say 
obviously imported because these worms are only found in 
cultivated ground and near the coast ; as civilisation is left 
behind these species decrease and are replaced by the 
truly indigenous species. Among the twenty species of 
Acanthodrilus are included two or three which occur in the 
Falkland Islands and in South Georgia. Turning to New 
Zealand we find that out of twenty indigenous species nine 
are Acanthodrilus, six belong to the closely allied genera 
Octochcehis, Deinodrilus, and Plagiochcsta, three are Micro- 
scolex, while the two remaining are a Perichceta and a 
Megascolides, two genera which are eminently characteristic 
of the adjoining continent of Australia. Between New 
Zealand and South America is a long stretch of ocean, 
sparsely scattered over which are islands of volcanic origin. 
From three of these islands earthworms have been collected. 
In Kerguelen and Marion Island is a species of Acantho- 
drilus peculiar to those islands, and I have lately received, 
and am describing in the forthcoming June number of the 
Proceedings of the Zoological Society, a second species of 
that genus from Macquarie Island. The significance of 
these facts will be more apparent when we consider how 
far the genera that have been referred to in the fore- 
going are distributed outside of this antarctic area. Micro- 
scolex is found in many parts of central and the warmer 
western regions of North America ; it has been met with 
also in Europe, Algeria and Teneriffe. Acanthodrilus 
occurs in Australia where it is represented by three species, 
all of which however inhabit the eastern half of the island 
continent, that part in fact which is nearest to New 
Zealand ; Acanthodrilus has one species in Natal, one in 
New Caledonia and two in North America. 

We have evidently therefore a fauna of earthworms 
peculiar to the antarctic region, into which more northern 
forms have been able to make but slight inroads and from 
which but few stragglers have wandered. 


As to other distributional facts and theories, it is 
probable that I have underestimated in my Monograph the 
distinctness of the Palaearctic and the Nearctic regions of 
Mr. Sclater. I was disposed to unite them into one Hol- 
arctic as Professor Newton has called it. Further investi- 
gations have tended to emphasise the justice of separating 
these two regions. This evidence has been mainly collected 
by the industry of Dr. Gustav Eisen, of San Francisco ; 
but others whose names and memoirs will be found quoted 
in the list of literature at the end of this article have added 
details of importance. The North American continent is 
inhabited by a fair number of peculiar genera, of which 
Diplocardia, originally described some years since by 
Garman, has four species (partly referred to the undoubtedly 
synonymous genus Geodrilus) ; there are also peculiar to 
this region Phoenicodrilus, nearly related to the central 
and South America Ocuerodrilus, and Sparganophilus ; of 
this latter genus the original species was found by 
Benham in the Thames ; but as there are half a dozen 
American species it seems likely that its occurrence in 
England is a case of importation. Bimastos is a genus 
perhaps justly separable from Allolobophora, from which it 
chiefly differs in the large size (for a Lumbricid) of the 
glandular sac in which the efferent male ducts terminate. 
Besides these peculiar genera are a few species of the 
Central and South American genera Ocuerodrilus and 
Kerria, and of the almost world-wide Benhamia. Aleodrilus 
is an Acanthodrilid that Eisen is disposed to separate from 
Diplocardia ; two species of Acanthodrilus complete the 
list of non-European inhabitants of the North American 
Continent. But in addition to these are a number of Allolo- 
bophora and Lumbricus — the characteristic forms ol the 
Palaearctic region — two or three of which are, however, so 
far as our present knowledge goes peculiar to North 
America. These facts perhaps justify the retention of 
the Nearctic region, and they are perhaps also significant 
in that the peculiar forms are western in range — a possible 
indication of their approaching extirpation by European 
species introduced by commerce. 


The original indigenous forms, South American in 
character, may be regarded as having been gradually 
driven to the west by the encroachment of artificially in- 
troduced species. In other respects the geographical regions 
indicated by the distribution of earthworms agree fairly well 
with the generally received scheme of Mr. Sclater. The 
Ethiopian region is peculiarly distinct ; the Neotropical is 
also nearly if not quite as plainly marked ; but the Oriental 
fades into the Australian, and it is indeed not easy to 
separate them at all. 

The only other matter affecting the distribution of earth- 
worms with which I shall deal here is the question of 
oceanic islands. Our information upon the subject is not 
however by any means extensive ; the largest collection 
made is due to the energy of Mr. Perkins, and has been 
described by me in a paper communicated to the Zoological 
Society. These worms were gathered in the Sandwich Is- 
lands, and belong to a number of species of which only two 
(and a doubtful third) have not been found elsewhere ; 
these two belong to the genus Perichccta, a genus prevalent 
in tropical regions, especially of the old world. That the 
bulk of the species known from these and other oceanic 
islands are forms which have been in all probability intro- 
duced by accidental transference by man is rather what might 
be expected from the limited powers of independent travel 
possessed by these animals. There is at present no certain 
evidence that there are any truly indigenous earthworms in 
oceanic islands, with the exception of Kerguelen — a fact 
which as I have already hinted may be due to other causes. 

To Linnaeus only a single species of earthworm was 
known, his Liimbricus terrestris, now believed to have beeii 
a compound of more than one species. Grube in his 
Familie der Anneliden, published in 1851, reckoned up 
only forty-two earthworms, and of these one or two are 
now known not to be earthworms at all, and of the re- 
mainder many are unrecognisable or synonyms. Since that 
period the increase of new forms has gone on — of late 
with extreme rapidity ; at the present moment we are 
acquainted with rather over 500 distinct and well char- 


acterised species. And this estimate does not take into 
consideration subspecies or well marked varieties, and pays 
no attention to " species incertae ". Of aquatic Oligochceta 
150 is about the number of known species ; but this 
group is decidedly less known than the former. As with 
other groups of animals this great increase in the number 
of known species has added to our knowledge of anatomical 
fact, but rendered harder the formation of classificatory 
schemes. No indistinctness, however, has arisen to blur 
the perfectly sharp outlines of the group Oligochaeta, no 
''intermediate" forms have been discovered whose relega- 
tion to the group is a matter of uncertainty or convenience. 
At the same time a few approximations in structure to the 
leeches on the one hand, and to the Polychseta on the 
other have been discovered ; but these are in no case of 
first-rate importance. Perhaps the most remarkable is the 
description of the gills of the African genus Alma. This 
worm was originally described under that name by Grube 
in 1855. Thirty-four years later Levinsen, apparently in 
ignorance of Grube's paper, named a fragment of what was 
obviously the same worm Digitibranckus, and described in 
the same paper Siphonogaster, an Annelid characterised by 
a pair of long processes an inch or so in length, and of a 
spatula-like form arising from the eighteenth segment. 
These have been subsequently shown to be processes con- 
taining the outer section of the sperm duct which opens 
near to the extremity. Michaelsen showed that all these 
three worms are identical, and has thus been able to put 
beyond question the existence of a true earthworm l with 
branched retractile gills on the posterior segments of the 
body. It was not by any means clear from the earlier 
descriptions that the gilled worm was not a Polychaet. 
Among the lower aquatic Oligochsetes there are at least 
three gilled forms, apart from Dero which has a circlet 
of ciliated processes, with vascular twigs lying round the 
anus. These forms are Chcstobranckus of Bourne, and 
Branchiura and Hesperodrilus branchiatus of myself. In 

1 Structurally ; in habit it is aquatic. 


the two latter (which are allied to Tubifex) are contractile 
branchiae, not branched however, on some of the posterior 
segments of the body. More numerous are indications of 
affinity with the leeches. I may, in the first place, refer 
to that group of parasitic Oligochaeta, once placed among 
the leeches but now usually allowed to be true Oligochaeta, 
for which Vejdovsky has proposed the name of Disco- 
drilidae on account of their posterior sucker. An American 
genus Bdellodrilus has lately been studied with care by 
Moore whose results entirely confirm the placing of the 
worms amono- the Oligochaeta and their removal from the 
leeches. Their chief points of likeness to the Hirudinea are 
(1) absence of setae ; (2) existence of jaws ; (3) presence of a 
sucker ; (4) median unpaired character of reproductive pores. 
The first and last of these characters are, however, 
found in a few undoubted Oligochaeta, for instance, Anachczta, 
as its name denotes, has no setae, and besides Mr. Moore 
describes large gland cells in Bdellodrilus which may re- 
present setigerous cells of Oligochaeta. As to the median 
generative pores they are very frequent among Oligochaeta. 
The reproductive organs themselves are decidedly upon the 
Oligochaetous pattern. The gonads are entirely free from 
their ducts, and there is a single spermatheca, a structure 
entirely wanting among the true leeches. The male ducts 
are two pairs, opening freely by ciliated mouths into the 
coelom and uniting into a common terminal atrium. Their 
arrangement recalls that of the Lumbriculidae. The ovaries 
are proliferations of the coelomic walls and their contents 
escape to the exterior by a slit in the body walls lined by 
epithelium, a kind of rudimentary oviduct paralleled in the 
Enchytraeidae, and in the Eudrilid Nentertodrilus. There 
is nothing leech-like about the reproductive organs, except- 
ing the terminal penis — a structure, however, which is 
also found in many Eudrilids and in some other Oligo- 
chaeta. The conclusions of the author that the Disco- 
drilidae are Oligochaeta slightly modified for a parasitic 
life is quite borne out by their structure. We may admit 
at the same time that this modification is in the direction 
of the leeches. 



In addition to questions of relationship to other neigh- 
bouring groups, recent investigation has brought to light 
facts of interest in the anatomy of the Oligochaeta which 
bear upon the mutual affinities of the families and genera 
into which the order is divided. In this direction the main 
discoveries of importance relate to the excretory system. In 
all the simple aquatic genera each segment of the body 
contains a single pair of nephridia ; as a rule these organs 
are wanting in the anterior segments, and Professor Bourne 
was unable to find any nephridia at all in Uncinais littoralis. 
The absence of nephridia in the anterior segments of the 
body, however, also characterises certain earthworms. It 
was originally described by Perrier in Pontodrilus, and all 
the species of this genus (6) are in the same condition. 
More recently Benham and Risen have shown that the 
same state of affairs characterises the aquatic Geoscolecid 
Sparganophilus. A distinction therefore between the 
Limicolae and Terricolse of Claparede quite breaks down. 
That these genera have no gizzard or calciferous glands 
(or at most the rudiments of a gizzard) is evidence of general 
degradation, which may have something to do with their 
aquatic or semiaquatic existence. It suggests too that 
the simplification in structure of the Limicolae of Claparede 
may be rather due to degeneration than to the retention of 
primitive characters. 

Among the earthworms, however, the single pair of 
nephridia to each segment is far from being the rule. In 
a large number of genera the nephridia are multiple. Two 
pairs in each segment exist in BracJiydriliis ; three pairs in 
Trinephrus; and Eisen has lately shown that in certain 
North American Benhamias there may be three or four 
distinct and separate pairs each with its own internal funnel 
and external pore. The complexity of the excretory 
system culminates in Perichceta where a single segment may 
be furnished with probably at least one hundred external 
nephridiopores. It is, however, a question whether in 
this latter case there is really an intercommunication be- 
tween the several nephridia of each segment, and between 
those of adjacent segments as has been alleged by Spencer 


and myself. The matter requires renewed investigation. 
In any case Bourne, Vejdovsky and I have shown that the 
" plectonephric " condition, as Benham has termed these 
diffuse nephridial tubes, is preceded by a series of paired 
nephridia one pair to each segment. This has been proved 
in Pericktzta, Qciochcetus and Megascolides. The nephridium 
elongates and becomes thrown into loops, each loop finally 
appears in Megascolides to break away and to form a 
distinct and separate nephridium. It is clear, therefore, 
that whether or not the connection is retained in Octochcetris 
and Perichceta there is originally a connection, so that that 
matter is of less importance than the alleged intercom- 
munication from segment to segment. This multiple 
arrangement of the nephridia is only found in the families 
Acanthodrilidae, Perichsetidae and Cryptodrilidae, and is the 
principal argument for uniting them into one superfamily, 
Megascolicides, as I have done in my Monograph. Brachy- 
drihts, however, is a member of the family Geoscolicidse, 
but it has only two pairs of nephridia to each segment ; 
there is nothing like the complicated system of Perichceta. 
This family Geoscolicidse has been through the recent re- 
searches of Rosa and Michaelsen brought still nearer to the 
Lumbricidse. It was always difficult to separate them, 
mainly on account of the aquatic Criodrilus, now it is 
practically impossible unless we accept Michaelsen's inter- 
mediate family Criodrilidse. The ornament setae which 
used to be a distinctive mark of the Geoscolicidse have 
been found by Michaelsen in Allolobophora moebii and 
in A. lonnbergi ; many Geoscolicidae, e.g., Microckceta are 
distinguished by the fact that instead of a single pair 
of spermathecae in each of those segments which con- 
tain them there are a considerable number of minute 
pouches ; this distinction, however, falls to the ground 
since more than one Allolobophora is now known to 
possess the same character — which has moreover been met 
with in Perichceta. It is in these two families that most 
instances are met with of total absence of spermathecae ; 
Kynotus, a Madagascar genus, is anteclitellian like the 
Lumbricidae, and in short it seems impossible to lay down any 


set of characters which should absolutely separate the two 
families. Several members of the two families are aquatic ; 
thus among the Geoscolicicke Bilimba (with which Michaelsen 
now suggests to unite Horst's A nnadri/us and Glyphidrilus), 
Criodrilus, whose range the same author has lately ex- 
tended to South America, Alma and Sparganophilus. Of 
Lumbricidae Allurus is the only form which is often 
aquatic. Michaelsen has dwelt upon the fact that all of 
these, with the exception of Sparganophilus, have the 
body generally or at least the posterior region markedly 
quadrangular in outlines with the setae implanted at the 
four corners. This is an apparent consequence or at least 
concomitant of aquatic life which is more curious than 
explicable. So much then for recent modifications of the 
systematic arrangement of the group. I shall deal finally 
with various anatomical and histological discoveries which 
have a general interest unconnected with systematic rela- 
tions. The most important work under this heading is 
undoubtedly the recent investigations into the structure of 
the remarkable family Eudrilidae, a well-defined family 
whose boundaries have not become in the least indistinct 
by the discovery of new forms. The family is remarkable 
on account of its distribution as well as on account of 
certain anatomical peculiarities. It is limited to tropical 
Africa — to the Ethiopian region of Sclater, with the sole 
exception of the type genus Eudrilus, whose ubiquitous- 
ness, however (America, West Indies, India and the East 
generally, New Zealand, etc.), makes one suspect direct 
transference by man. This family is chiefly interesting on 
the anatomical side by reason of the illustration which it 
gives of two phenomena, viz., substitution of organs and 
change in function of organs. 

In all Oligochaeta the ovaries are paired (rarely 
unpaired) structures which arise from the peritoneal 
epithelium of the earthworms invariably the thirteenth 
segment. They are totally unconnected with the oviducts 
whose open mouths are placed exactly opposite to them. 
In the Eudrilidae these gonads are enclosed in sacs 
which communicate with a system of sacs the complexity of 


which varies in different genera, and of which it would be 
impossible to give any detailed account without the assis- 
tance of figures. There is a separate receptaculum ovorum 
like that of the common earthworm, with which is connected 
the oviduct. This system of sacs, through which the ova 
can travel in so far as there are no physical hindrances, also 
contain sperm, and play the part of spermatheca? or a sperma- 
theca. They commonly open by a single ventral pore ; 
sometimes the structures are paired as in the genus Eudrilus 
itself. Now these pouches generally contain sperm, and 
there is therefore the possibility of the ova being impreg- 
nated within them ; Michaelsen has even suggested that 
some species are viviparous. In a few genera, for example 
in Heliodrilus, these pouches do not communicate with the 
exterior except through the oviducts. They appear to do 
so by a large ventral pore, but when careful sections are 
made it is found that this pore is the mouth of a closed sac, 
exactly like a spermatheca, which is enclosed within the 
large pouch. Thus the ccelomic nature of this system of 
sacs is established on anatomical grounds, and develop- 
mentally they have been shown, at least in one genus, to 
be derivatives of the intersegmental septa just as are the 
sperm sacs of other earthworms ; their cavities are therefore 
separated portions of the general ccelom. But, as already 
mentioned, in most cases they do open on to the exterior 
directly by a conspicuous orifice, and contain sperm which 
probably finds its way into them by this orifice. The fact 
that in some cases these sacs contain structures which are 
precisely like the spermathecae of other earthworms, and 
that in other cases where they open directly on to the 
exterior the character of the lining epithelium changes near 
to the orifice, becoming distinctly columnar, suggests that 
we have to do here with the substitution of sacs derived 
Irom the septa for the true spermathecae which are gradually 
disappearing, only the extremity being left in the majority 
ot cases. The second point with which I wish to deal 
concerns the calciferous glands. Most, but by no means 
all, earthworms possess one or more pairs of these organs, 
which are attached to and open into the cesophagus. What- 


ever may be their functions they contain crystals of car- 
bonate of lime, and have a rich vascular supply, the lining 
epithelium being much folded and therefore extensive. In 
some Eudrilidae these structures are absent or rather are so 
altered that they are nearly unrecognisable as calciferous 
glands. At the same time they have become more numerous. 
The structure is altered in that instead of an extensive lumen 
produced by the folding of an excretory epithelium there is 
a very short sac connected with the oesophagus, which 
is, however, enveloped by an extensive coating of cells 
which I regard as ccelomic cells, and among which meander 
abundant blood-vessels. These ccelomic cells, where they 
abut upon blood-vessels, very often lose their oval or 
rounded form and become columnar and at the same time 
more darkly staining. They surround the blood-vessel as if 
it were the lumen of a secreting gland, the cells themselves 
having acquired the appearance of a secreting epithelium. 
These phenomena suggest that we have to do here with a 
change of function on the part of the calciferous glands ; that 
their function of producing carbonate of lime, that their 
connection with alimentation has disappeared or is dis- 
appearing, and that a new function more intimately connected 
with the vascular system has supervened. There is a 
certain analogy here with the vertebrate liver which has 
certainly more functions than that of pouring bile into the 
intestine, though originally it may have been merely an 
annex of the alimentary canal. 

In histology there is only one matter to which I shall 
direct the attention of the reader. It concerns the minute 
structure of muscular fibres in the Oligochoeta. The careful 
researches of Cerfontaine have established the fact that the 
Oligochseta, like the leeches, have muscular fibres which 
consist of an outer sheath often radiately striated, the 
muscular substance, and a soft central core. Hesse, how- 
ever, while admitting this, goes a step further and 
endeavours to prove a resemblance to the muscular fibres 
of the Nematoidea. He figures in the Enchytraeidae and 
in the Lumbricidse a gap in the sheath of the fibre through 
which the soft less-modified protoplasm of the interior com- 


municates with a pear-shaped nucleated body outside. If 
these observations prove ultimately to be correct it is clear 
that there is a close resemblance in this particular between 
the Oligochseta and the Nematoidea. 


Beddard. A Monograph of the Order Oligochseta, Oxford : 

Clarendon Press. 
ElSEN. Pacific Coast Oligochaeta. Mem. Calif. Acad. Set., vol. ii., 

Hesse. Beitrage zur Kenntniss des Baues der Enchytraeiden. 

Zcitschr. fur iviss Zoo/., 1893. 
HESSE. Zur vergleichenden Anatomie der Oligochaeten. Ibid., 

MlCHAELSEN. Zur Kenntniss der Oligochaeten. Abh. Nat. Ver., 

Hamburg, 1895. 
H. F. MOORE. On the Structure of Bimastos palustris. Journ. 

Morph., 1895. 
J. P. MOORE. The Anatomy of Bdellodrilus illuminatus. Ibid. 
ROSA. Allolobophora dugesii. Boll. Mus. Zooi, Torino, 1895. 
BOURNE. In Quart. Journ. Micr. Sci., 1894. 
SMITH. Notes on Species of North American Oligochaeta. Bull. 

Illinois State Lab. 

F. E. Beddard. 


IN a former article 1 a sketch of the state of our know- 
ledge as to the relative atomic weights of hydrogen and 
oxygen was given. It was there shown that although the 
great mass of the evidence was in favour of the atomic 
weight of oxygen being about 15*88 times that of hydrogen 
yet there was a certain amount of experimental work by 
well-known and tried observers which seemed irreconcilable 
with this result, the chief paper (1) being that of Professor 
Julius Thomsen of Copenhagen, and based on the propor- 
tion by weight in which ammonia and hydrochloric acid 
combine to form neutral ammonium chloride. In a short 
paper by the late Lothar Meyer (2) it was proved con- 
clusively how little value could be attached to a determina- 
tion of this nature however accurate and careful the mani- 
pulative work might be. 

Any hopes which might have survived in the minds of 
the most ardent follower of Prout, that the atomic weight 
of oxygen is exactly sixteen times that of hydrogen, must 
now be dispelled by the recent publications of E. W. 
Morley (3) and of Thomsen (4) himself. The work of 
Morley is so conclusive, and has been carried out with 
such untiring patience and skill, that to any one who reads 
the clear account which he gives of his methods and of the 
various checks employed, it must be quite evident that that 
type of worker of whom we regard Stas as the chief is not 
yet extinct, in spite of the prevailing view that one must 
publish as many papers as possible in the least possible 
time before one can be said to engage in " original re- 
search ". Morley's scheme for the complete determination 
of the relative atomic weights of oxygen and hydrogen is 
a most ambitious one, and, although his results are quite 
conclusive now, it is much to be regretted that bad health 
and other circumstances over which he had no control 
(such as a workman pushing a brick through a wall on to a 

1 August, 1894. 


delicate piece of glass apparatus) have up to the present 
time prevented him from carrying out his original pro- 
gramme in its entirety. 

The paper consists of four distinct parts — 

I. The determination of the weight of a litre of oxygen. 

II. The determination of the weight of a litre of 


III. The ratio by volume in which these two gases 

combine to form water. 

IV. The synthesis of water from known weights of 

hydrogen and oxygen, the weight of the water 
formed being also accurately determined. 

It would be impossible to give any idea of the precau- 
tions taken to obtain results free from all objections in a 
sketch so short as this must be, for such details the 
original memoir must be consulted ; only a summary of 
the results obtained can here be given. 

Three methods were adopted to determine the weight 
of a litre of oxygen. In the first method the barometer 
and thermometer were used, and the gases weighed in 
balloons holding in three of the experiments about 9 litres, 
and in the other six about 21^ litres. 

In the second method a globe of pure and dry hydrogen 
was used as the standard for temperature and pressure, the 
globe containing the oxygen having its pressure deter- 
mined at the same temperature as that of the hydrogen 
by means of a very sensitive differential manometer. 

In the third method the globes were filled with oxygen 
when they were immersed in melting ice and the pressure 
accurately determined at the moment of closing. This 
method had the disadvantage of wetting the surface of the 
globes, and probably thereby changing their weight (although 
this was duly investigated). 

The values obtained by these three methods for the 
weight of 1 litre of oxygen under normal conditions of 
temperature and pressure at sea level in lat. 45 were 

By use of thermometer and manometer- = 1-42879 + "000034. 
By compensation - 0=1-42887 + -000048. 

By use of ice and barometer - - = 1*42917 + -000048. 


From various considerations taking into account errors 
incidental to certain methods and liability to constant errors 
Morley gives the most probable value as i "42900 ±0*000034. 

In the same way experiments were made with hydrogen 
and in live series but practically by three methods. 

First method was practically the same as the first series 
of oxygen experiments. 

Second method was like the third oxygen series. 

Third method utilised the power of absorbing hydrogen 
possessed by palladium. The hydrogen was weighed in 
the palladium and expelled into globes, and its volume and 
pressure determined at the temperature of melting ice. 
Series III., IV. and V. were made by this method, but 
the apparatus employed varied somewhat in the various 

The values which result from these experiments are 

Series I. D h = -089938 gram. 

Series II. D h = '089970 gram. 

Series III. D h = -089886 + -0000049 gram. 

Series IV. D h = -089880 + -0000088 gram. 

Series V. D h = -089866 + -0000034 gram. 

The higher results of Series I. and II. are possibly due to 
some constant error, probably traces of mercury vapour. 
The most probable value is 

D h = -089873 + 0*0000027 gram. 

Part III. of the paper begins with a sketch of the methods 
it was proposed to employ to determine the volumetric 
composition of water. Of the three methods proposed Morley 
unfortunately has only been able to carry out the one which 
is the least satisfactory, viz., the determination of the 
density of electrolytic gas and of the excess of hydrogen 
over and above what the oxygen can unite with. Leduc 
made a similar density determination, but apparently 
assumed that the hydrogen and oxygen were in the exact 
proportions in which they would recombine to form water. 
Morley found that he always had an excess of hydrogen 
when he kept his voltameter in ice and water, but that 
when the temperature was allowed to rise to about 20° C. 
then oxygen was in slight excess, so that no doubt at a 


certain temperature the gases do come off in atomic propor- 
tions. In each experiment the weight of the gases given 
off was about 23 grams. 

The weight of a litre of the gas thus given off from 
solution of soda made from clean sodium was — 

°'5355 1Q ± o-ooooio, 
and corresponds to a mixture of one volume of oxygen with 
2*00357 volumes of hydrogen, but the excess of hydrogen 
was found to be - ooo88 giving therefore the ratio in which 
the gases combine as 1 : 2*00269. 

Part IV. gives an account of experiments in which 
hydrogen was weighed in palladium foil, oxygen was 
weighed in a globe, these were then made to combine, and 
the water produced was weighed also. 

From these experiments we get the following values for 
the atomic weight of oxygen : — 

(1) From the ratio of hydrogen and oxygen, - - 15-8792 

(2) From the ratio of hydrogen and water, - - - 15-8785 

or as a mean, ------ 15879 

From Parts I., II., III. of the memoir we get 

1*42000 2 

— 1 7 ~ x = 15-879 

•089873 2-00269 

How excellent Morley's work is can perhaps best be 
seen by comparing his results with the means of those of 
previous experimenters, 


summary. Morley. 

Density of oxygen at Paris, - - 1-42961 1-42945 

Density of hydrogen at Paris, - - -08991 -089901 

Ratio of densities mean of all previous determinations, - - 15-9005 
Ratio of densities, Morley's, 15-9002 

Ratio of combining volumes, Morley, - 2*00269 
„ ,, Scott, - 2*00285 

,, ,, Leduc, - 2*0037 (corrected = 2 0024) 

Although the results obtained by Thomsen agree 
wonderfully well with those of Morley it is not because his 
apparatus and his methods of working are so carefully 
elaborated. On the contrary what strikes one most forcibly 
is the extreme simplicity of the apparatus and mode of 


working it as well as the neglect of certain precautions 
which could well have been taken, and ought to have been 
taken in an attempt to settle such an important constant as 
the present ; such precautions as to weighing with counter- 
poises of equal volume, for example, seem to have been 

The method was to determine, firstly, the weight of 
hydrogen given off from unit weight of aluminium when 
dissolved in strong potash solution ; secondly, by supplying 
oxygen to a small combustion chamber so as to burn the 
hydrogen evolved from a known weight of aluminium, and 
collect all the water formed in the apparatus, one gets thus 
the gain of the equivalent amount of oxygen to the hydro- 
gen and to the aluminium. The only corrections not of 
the simplest order were due to the oxygen and hydrogen 
remaining in the apparatus or which had to be evolved after 
the combustion had ceased. It was not found possible to 
burn all the hydrogen evolved completely as the current 
became so very slow when a very little aluminium remained 
undissolved. The aluminium did not require to be perfectly 
pure as long as it gave off no other gas than hydrogen. It 
was found that 162 "3705 grams of aluminium gave off 
1 8*1778 grams of hydrogen giving the ratio 


p — = 0*111902 + "000015 


as the mean of twenty-one experiments. 

The weight of oxygen required to combine with 86*9358 
grains of aluminium (or rather with the hydrogen evolved 
by its solution in potash) was found to be yyiSyG grams 
from which we get the ratio 

°*W^ = -88787 ±00001 S 

from which two results we get 

O -88787 

5- = — — = 7 '9345 

Ho '11190 

or — = i?-86qo + "oo22 
H D ~ 

We seem to have every reason now to regard it as com- 
pletely proved that the atomic weight of oxygen is 15*87 to 


r 5*88 times that of hydrogen, the higher value being in all 
probability the more correct. 

Having now satisfactory determinations of our funda- 
mental ratio we still require other ratios to be able to de- 
termine conveniently the atomic weights of many elements. 
If an element forms many compounds with oxygen it is 
never safe to conclude without the most rigorous proof that 
we have a pure oxide absolutely free from the other oxides 
of the same element. Hence determinations of atomic 
weights made by the reduction of oxides to the element or 
of one oxide to a lower one or of the oxidation of an 
element to an oxide or of one oxide to a higher oxide must 
always be accepted with caution. The use of the haloid 
compounds (especially those of bromine), of many elements, 
is of the greatest value, and for this we require an exact 
knowledge of the ratio bromine : oxygen. For this we 
depend chiefly on the classical work of Stas. The publica- 
tion of the complete works (5) of J. S. Stas under the able 
editorship of Professor W. Spring, of Liege, enables every 
one now to obtain in an elegant and convenient form these 
models and masterpieces of accurate research which were 
formerly so difficult to procure. How great the contrast 
between the work of Stas and too much of that turned out 
at the present day a glance at almost any page of his 
works will show. Every step was proved most conclusively, 
however simple and even axiomatic it may seem to us now, 
before he proceeded to more elaborate propositions and 
deductions. For instance, in his Nouvelles Recherches he 
begins by proving that ammonium chloride prepared from 
absolutely different sources and purified in different ways 
always contains exactly the same proportion of chlorine, 
and that the same weight of each sample precipitated 
exactly the same amount of silver from its solution in 
nitric acid. He obtained his ammonia from ordinary sal 
ammoniac after destroying any organic bases by a treat- 
ment with aqua regia, and from commercial ammonium 
sulphate by a similar purification, by heating it to a high 
temperature with strong sulphuric acid, and then oxidation 
with nitric acid, and from potassium nitrite by reduction in an 


alkaline solution with purified metallic zinc. The ammonium 
chloride was sublimed now in a current of ammonia gas, 
now in vacuo, but the results obtained showed that for the 
complete precipitation of 100,000 parts of silver, 49,592 to 
49,602 parts of ammonium chloride were required. In 
other words, the extreme difference in a large number of 
determinations carried out with very considerable modifica- 
tions only amounted to one part in five thousand. 

Having thus proved that a compound always contains 
the same proportion of its constituent elements it was 
essential for his purpose as well as for the complete 
establishment of the atomic theory to prove that the equiva- 
lent weight of an element was not affected in the slightest 
degree by the various elements with which it might 
combine. To take an example, silver combines with 
iodine to form the iodide, and with iodine and oxygen to 
form the iodate, and these compounds are represented by 
the formulae Agl and Ag I0 3 respectively. It was just 
possible, one might even say probable, that the ratio of silver 
to iodine in the one compound might not be the same as 
that in the other, but that it would be modified by the large 
quantity of oxygen present in that other substance. If, 
however, the elements consist of small particles alike in all 
respects, such a variation would be impossible, and the 
relative masses of silver and iodine in the iodide and in the 
iodate must be absolutely the same. To prove this may 
seem very easy, but Stas found it by no means so, lor 
whenever he prepared his silver iodate by precipitation from 
the nitrate, after the reduction with sulphurous acid there 
was always a small excess of silver over and above the 
iodine present. This he finally traced to a minute quantity 
of the nitrate being carried down mechanically by the 
iodate, but so firmly held that no amount of washing would 
remove it. By using other soluble salts of silver such as 
the sulphate and the dithionate, however, he was able to 
prepare silver iodate so pure that on reduction to silver 
iodide not the slightest trace of either silver or iodine re- 
mained in excess. In the case of that prepared from the 
nitrate the excess of silver only amounted to one part in 


3,000,000. These simple experiments give us some idea 
as to how hard it is to obtain even very simple compounds 
in a state of absolute purity. Having thus laid the founda- 
tions for his further work, and shown that the combining 
proportions of elements are mathematically exact, Stas con- 
sidered no labour too great if thereby he could obtain more 
accurate values for these proportions. Any work done 
since his determinations has only tended to uphold his 
values and to increase our admiration for his work. 

The great value of very accurate experimental work 
has been most strikingly exemplified by Lord Rayleigh's 
determinations of the density of nitrogen (6). He 
found that the nitrogen which he could obtain from air 
alone by removing the oxygen was very little denser, but 
was always denser than that prepared from the air with the 
aid of ammonia by Harcourt's method, and that the nitro- 
gen prepared from ammonia or from any compound had 
always the same density, and that this was still lighter than 
that partly from air and partly from ammonia. From this 
he concluded that besides nitrogen the atmosphere must 
contain another constituent still denser, which like nitrogen 
resisted the action of iron and copper as well as their oxides, 
even when very strongly heated. By combining the 
nitrogen with oxygen after the method of Cavendish, or by 
causing the nitrogen to unite with metallic magnesium, a 
new gas to which the name of argon has been given was 
finally separated by Rayleigh and Ramsay after much 
laborious work. The detection in the atmosphere of a 
constituent hitherto unsuspected as well as its isolation are 
apparently only the first fruits of a number of more or less 
startling discoveries Mowing directly from Lord Rayleigh's 
very accurate work. The molecular weights of argon 
(7) and helium (8) are respectively 40 and 4, and if their 
molecules are monatomic this would give us the same 
numbers for their atomic weights, but if the molecules are 
diatomic, as is probable, these numbers would be halved for 
the atomic weights. It is far from certain that either what 
we call argon or what we call helium is not a mixture of 
several similar substances. 


Several atomic weights have been redetermined with 
great care, and of these determinations perhaps those 
of T. W. Richards of barium and of strontium are the most 
accurate and most interesting. By an exhaustive research 
on barium bromide he deduces the value Ba = 137*434 
(O = 16) (9). From a similar study of barium chloride the 
value Ba = 137*440 is deduced (10). 

This value is notably higher than that usually accepted 
and is no doubt due to the careful elimination of small 
quantities of strontium and calcium which have contaminated 
the preparations of earlier experimenters. From a study of 
strontium bromide Richards found Sr = 87*659 (O = 16) 

Still more recently the atomic weight of zinc has been 
determined by Richards and Rogers again by means of the 
bromide and precipitation with silver, and as a mean they 
find the value (Zn = 65*404) (O = 16) (12). 

In all the above determinations Richards estimated the 
percentage of silver in his haloid silver salt and showed it 
to be identical with that found by Stas, thus placing his 
work on the same footing and guaranteeing in this way its 
very high accuracy. 

In 1888 two other American experimenters, Burton and 
Morse (13), published the results of their work on the same 
atomic weight which they arrived at by means of the con- 
version of the metal into the oxide by treatment with nitric 
acid and ignition of the nitrate. Although their work 
agrees throughout very well the value found is lower than 
that of Richards, due no doubt to the retention by the 
oxide of oxides of nitrogen as Marignac pointed out. In 
defending their work against this objection they expose 
their want of knowledge of the commonest reactions in such 
a way as to make one distrust all their work. The perusal 
of their paper provides much food for reflection of a serious 
nature although it does give a certain amount of instruction 
as well as amusement. They carry out their weighings to 
*ooooi of a gram and pretend to detect differences of this 
minute amount in porcelain crucibles which have been 
heated up to the melting point of steel. In their account 


of the purification of metallic zinc by distillation in vacuo it 
is rather odd to find it stated that indiarubber tubing with 
glycerine joints could not be used because the vapours of zinc 
and of glycerine interact. What pressure of the vapour of 
each is likely to exist at the highest temperature to which the 
joints would ever be subjected ? The presence of gold in 
the nitric acid distilled from a platinum still, and coming 
from the gold solder used in it sounds also rather peculiar. 
One knows that very finely divided gold will dissolve in 
fuming nitric acid if kept cold, but one could hardly have 
thought of finding it as an impurity in nitric acid prepared 
by distillation. But the gem of all the statements comes at 
the end of the paper when these two rising experimentalists 
proceed to criticise Marignac's work (14), and finally to 
teach him and us how we ought to test for oxides of nitro- 
gen by means of starch and potassium iodide. After 
proving to their own satisfaction by a process which cannot 
reveal the presence of any of these oxides that they are 
therefore obviously absent, they conclude that Marignac 
was ignorant of the necessary precautions which must be 
taken to exclude oxygen, especially that of keeping the 
solution practically boiling so that the steam may keep out 
the air. It is usually accepted as a well-established fact 
that the delicacy of this reaction decreases rapidly with rise 
in temperature, and that the colour goes completely before 
the boiling point is reached, even in the presence of 
relatively large quantities of free iodine. 

Amongst other noteworthy determinations of atomic 
weights made recently are those of Winkler, who finds 
the values Ni = 58*91 and Co = 59*67 by means of the 
reaction between the chlorides and silver (15); and still 
more recently Ni = 5871 and Co = 59*37 (16) by deter- 
mining the amount of iodine required to unite with the 
pure metal. Winkler uses the value Ag = 107*66, if we 
use O = 16 or Ag = 107*93 these last values become 

Ni = 58-863 
Co = 59*517 

The determinations of the atomic weight of boron by 



Ramsay and Acton (17), as well as by Rimbach (18), are 
very interesting as examples of various methods of attacking 
this problem, and which give fair results, but they can 
hardly be said to have given results possessing greater 
accuracy than those of Abrahall (19). 

Of all the elements of which the atomic weights are still 
in doubt, and of which the determinations are very unsatis- 
factory, by far the most interesting is undoubtedly tellurium. 
According to the periodic classification of the elements it 
ought, as is well known, to have an atomic weight less than 
that of iodine, but all the most satisfactory determinations 
are irreconcilable with this, and make the atomic weight 
notably higher than that of iodine. The experiments made 
in recent years both by Brauner (20) and by Wills (21) 
agree in this, no matter what method is adopted as long as 
it is one which gives concordant results. The latest deter- 
minations, those of Staudenmeier (22) which start from 
telluric acid, give, according to him, the values 127*6, 
127*1, and 127*3 f° r three series of experiments in which 
different ratios were determined. He takes as his standard 
O = 16 and H = 1*0032. Staudenmeier upholds that tel- 
lurium is an element in opposition to Brauner who at one 
time maintained that it was a mixture of true tellurium with 
a higher homologue, but now concludes that this is very im- 
probable, and since the discovery of argon suggests that 
the assumed impurity may be a homologue of argon. 
Speculations of this nature are strongly to be discouraged 
and condemned, especially when their basis is nothing 
more than the assumed abnormality in the periodic ar- 
rangement of the elements coupled with a very decided 
want of agreement in the results of an experimenter's own 
work obtained by different methods. They may afford an 
easier way out of a difficulty than by working steadily at 
the causes of such discrepancies, but afford at best but a 
feeble and undignified cover for one's retreat. 

P.S. — About the middle of last month, and after the 
above article was written, Thomsen (23) published the 
results of some new determinations of the densities of 
oxygen and hydrogen. The oxygen was prepared by 


heating a mixture of potassium chlorate and ferric oxide, 
and the hydrogen from a solution of caustic potash by the 
action of metallic aluminium. The values found were : — 

Weight of one litre of oxygen at 0° C. and 760 mm. pressure, at 
sea-level in Latitude 45° - - - = 1*42906 grams. 
And of hydrogen similarly = -089947 gram. 

From these he deduces the ratio of the volumes in which 
they must combine to form water to be 1 : 2*00237. 


(1) Thomsen, J. Experimentelle Untersuchungen zur Feststel- 

lung des Verhaltnisses zwischen den Atomgewichten des 
Sauerstoffs und Wasserstoffs. Zeitschrift fur physikalisdie 
Chemie, xiii., 398, 1894. 

(2) Meyer, L., und Seubert, K. Ueber das Verhaltniss der 

Atomgewichte des Wasserstoffes und des Sauerstoffes. 
BericJite ler deutsclien chemischen Gesellschaft, xxvii., 2770- 
2773, 1894. 

(3) Morley, E. W. On the Densities of Oxygen and Hydrogen, 

and on the Ratio of their Atomic Weights. Smithsonian 
Contributions to Knoivlcdge, No. 980, 1895. 

(4) THOMSEN, J. Experimentelle Untersuchung iiber das Atom- 

gewichts verhaltniss zwischen Sauerstoff und Wasserstoff. 
ZeitscJirift fur anorganische CJiemie, xi., 14, 1896. 

(5) STAS, J. S. Giuvres completes. Edited by Professor W. 

Spring, Bruxelles, 1894. 

(6) Rayleigh, Lord. On an Anomaly Encountered in Deter- 

minations of the Density of Nitrogen Gas. Proceedings of 
the Royal Society, lv., 340, 1 894. 

(7) Rayleigh and Ramsay. Argon, a New Constituent of the 

Atmosphere. Philosophical Tra?isaction±, clxxxvi., A. 223, 

(8) Ramsay, W. Helium a Constituent of Certain Minerals. 

Journal of the CJiemical Society, Ixvii., 684, 1895. 

Langlet, N. A. Ueber das Atomgewicht des Heliums. 

Zeitschrift fur anorganische Chemie, x., 289, 1895. 

(9) Richards, T. W. A Revision of the Atomic Weight of 

Barium ; the Analysis of Baric Bromide. Proceedings of 
the American Academy of Arts and Sciences, xxviii., 1-30, 


(10) RICHARDS, T. W. A Revision of the Atomic Weight of 
Barium; the Analysis of Baric Chloride. Proceedings of the 
American Academy of Arts and Sciences, xxix., 55-91, 1893. 

(n) RICHARDS, T. W. A Revision of the Atomic Weight of 
Strontium ; the Analysis of Strontic Bromide. Proceedings 
of the American Academy of Arts and Sciences, xxx., 369- 
389, 1894. 

(12) RICHARDS, T. W. and ROGERS, E. F. Neubestimmung des 

Atomgewichtes von Zink ; analyse von Zinkbromid. Zeit- 
schrift fiir anorganische CJiemie, x., 1-24. 

(13) Morse, H. N. and Burton, W. M. The Atomic Weight 

of Zinc as Determined by the Composition of the Oxide. 
American Chemical Journal, x., 31 1-32 1, 1888. 

(14) Marignac,C.DE. Verification de quelques poids atomiques : 

Zinc. Archives des sciences physiques et naturelles [3] x., 193, 

(15) Winkler, C. Ueber die vermeintliche Zerlegbarkeit von 

Nickel und Kobalt und die Atomgewichte dieser Metalle. 
Zeitschrift fiir anorganische Chemie, iv., 10 and 462, 

(16) Winkler, C. Die Atomgewichte von Nickel und Kobalt. 

Zeitschrift fiir anorganische Chemie, viii., 291, 1895. 

(17) RAMSAY, W. and ASTON, E. The Atomic Weight of 

Boron. Journal of the Chemical Society, lxiii., 207, 1893. 

(18) RlMBACH, E. Zum Atomgewicht des Bors. Berichte der 

deutschen chemischen Gesellschaft, xxvi., 164, 1893. 

(19) ABRAHALL. The Atomic Weight of Boron. Journal of the 
Chemical Society, Ixi., 650-666, 1892. 

(20) BRAUNER, B. Experimental Researches on the Periodic 

Law. Journal of the Chemical Society, lv., 382-411, 1889. 

(21) Wills, W. L. On the Atomic Weight of Tellurium. Journal 

of the Chemical Society, xxxv., 704-713, 1879. 

(22) Staudenmeier, L. Untersuchungen iiber das Tellur. 

Zeitschrift fur anorganische Chemie, x., 189, 1895. 

(23) Thomsen, J. Experimentelle Untersuchung iiber die Dichte 

des Wasserstoffes und des Sauerstoffes. Zeitschrift fib 
anorganische Chemie, xii., 1-1 5, 1896. 

Alexander Scott. 




IT is clear that the theory of polystely forms an 
integral part of the general stelar doctrine, and we 
can hardly refuse to accept its main idea. But though each 
stele in the polystelic stems of, for instance, Aiiricula Ursi 
and many Polypodiacea^ is clearly the equivalent of the 
whole cylinder in the hypocotyl of the same plants, cases 
exist in which we seem forced to consider as steles, 
vascular strands which have none of the characters of the 
cylinder left about them. 

Deriving our idea of the typical stele from the mono- 
stelic organ, we come to consider it as essentially cylindrical 
and radially symmetrical. It is true that diarch roots are 
bilateral in structure, and the primary root and hypocotyl of 
very many ferns being diarch the steles of a great number 
of their stems are likewise diarch and hence bilateral. And 
this bilaterality often extends to the shape of the stele which 
becomes oval or band-shaped instead of circular in transverse 
section, the two protoxylems being situated at the extremi- 
ties of the figure. Another step is for the stele to become 
more or less semilunar in transverse section, so that it is no 
longer symmetrical about the plane passing through the 
protoxylems, but only about the bisecting plane perpendi- 
cular to this. And further the protoxylems may lose their 
symmetrical arrangement, or one only may be present, and 
this may be excentrically placed (Angiopteris). We clearly 
could not tell that such strands were steles if we had no 
knowledge of their connexions and disposition. At least 
as far as tissue arrangement goes they may often be said to 
have lost those characters which entitle them to the name. 
A similar difficulty meets us in the case of the vascular 
strands in many fern leaves. Undoubted steles found in 


the petiole, after repeated branchings gradually lose the 
phloem from their upper sides, and thus come to possess the 
collateral structure of the bundle of a Phanerogamic leaf. 
On the other hand the curved bundle in the petiole of 
Osmunda is certainly a meristele, if we may judge from its 
connexion with the bulky central cylinder of the mono- 
stelic stem, yet it is surrounded by a complete mantle of 
phloem, and indeed conforms in structure to many true 
steles (cf. 1 8). We may probably draw the same con- 
clusion as to the " petiolar steles " of Gleicheniacese (19). 

Similar facts appear to obtain in the polystelic genera of 
Phanerogams, upon which we may expect much new light 
from as yet unpublished researches. One instance is, 
however, too instructive to be omitted. A number of 
distinct steles arranged in a circle enter the peduncle of 
Auricula Delavayi (8, p. 304), fuse laterally, and become 
indistinguishable from a monostele, the central extra-stelar 
tissue passing over into pith. 1 Van Tieghem warns us (10, 
p. 768) not to confound such a structure formed in an 
essentially polystelic stem with an essentially monostelic 
stem. But if this sort of thing may occur, what guarantee 
have we that an "essentially monostelic" stem is really 
essentially monostelic, or, for the matter of that, that an 
" essentially polystelic " stem is really essentially polystelic ? 
If a stele can become a collateral bundle in the course of a 
shoot system, the same transformation may very well occur, 
or a collateral bundle may become a stele, in the course of 
descent ; at least we are quite debarred from dogmatically 
drawing or denying homologies between the one and the 
other. Of course we can speculate, and in some cases 
claim a fair degree of probability for our speculations, 
especially when we have a minute knowledge of all the 
facts in the anatomy of a given group, but since it is impos- 
sible to draw a sharp line between a stele and a vascular 
strand that is not a stele we are clearly not on very firm 
ground. There is certainly nothing to surprise us in this ; 

1 A similar state of things appears to obtain in some of the Palm roots 
investigated by Mr. Cormack. 


the instructive fact is that "there's such divinity doth 
hedge " a stele — indeed any morphological conception, as in 
almost every fresh case to prevent for a time our realisation 
of the truism that " Nature knows no sharp boundaries ". 
In the stelar doctrine, we have, no doubt, a classification 
that enables us to perceive a little more closely the direc- 
tions along which the various types of vascular system in 
the higher plants have been evolved, and that after all is 
the most we can expect. 



We have now to consider the developmental basis 
of the stelar theory. Let us take the Phanerogams first. 
It is well, as Dr. Scott (20) has already pointed out 
in this journal, to draw a distinction between de- 
velopment from the embryo, and development of the 
various axes from their permanently embryonic grow- 
ing points. It is clear, on reflection, that the former 
alone is comparable to ontogenetic development in animals, 
though it would be a mistake to suppose that the latter is 
not of importance to morphology. In the comparatively 
few types of monostelic plants with the anatomy of whose 
embryos we have a sufficient acquaintance, it appears 
that both in the plumule and radicle there is really a 
clear separation at the apex between central cylinder 
and cortex (plerome and periblem). But it is certainly 
open to doubt whether this distinction, as Hanstein 
thought, is really maintained at the growing points of the 
various axes throughout the life of the plant. Into the 
history of the differences of opinion on this point we need 
not enter. The inherent difficulties of arriving at valid 
conclusions from observations have been nearly as powerful 
as the subjective causes which have evidently influenced 
the views of the observers in creating the extraordinary 
discrepancies which exist between the various accounts. 

The method employed by Ludwig Koch (21 and 22), 
who recognised that the state of things at the growing 
point was likely to differ at different epochs of growth, and 


that hence conclusions drawn from observation of a few 
sections could not be final, marks a great advance on 
previous work. Koch claims to have proved (22), in Syringa 
and Berberis, that the single layer of cells immediately 
beneath the dermatogen, i.e., the periblem of earlier ob- 
servers, divides periclinally, during a period of leaf forma- 
tion, across the actual apex of the shoot, thus giving rise to 
three or four superposed layers of cells. It is clear that, 
if this is the case, all but the uppermost of these layers 
must become part of the plerome when the apex passes 
back to the state of possessing a single layered periblem. 
But though our author has convinced himself that this 
actually happens, his figures are not decisive. Most of the 
periclinal divisions which he shows in the periblem of the 
Lilac (Taf. xvi.) are clearly in connexion with the forma- 
tion of the leaf rudiments. In no case are such divisions 
shown across the actual apex. In fig. vi. periclinal walls are 
drawn in two periblem cells removed by one cell from 
the cell-group obviously concerned in the formation of a 
leaf rudiment, but these walls are also removed by one 
or two cells from the centre of the flat growing point, 
and considering how much this free surface is encroached 
upon by the developing- leaves {cf. fig. vii.) it is not 
at all clear that the periclinal wall in question is not 
precociously formed in a cell which will later be involved in 
the base of the leaf. Yet this single periclinal wall is 
really the sole evidence obtainable from his figures of the 
truth of Koch's view. Nevertheless the thorough method 
of investigation inaugurated by Koch must sooner or 
later settle the point. For the present we must admit 
that though Hanstein's case is made out for a certain small 
number of plants, the great majority of cases which have 
been investigated must remain doubtful. Van Tiegfhem 
(10, p. 776) does not definitely commit himself, though he 
implies the suggestion that Hanstein's three initial layers 
are universal in Phanerogams, though often not distinguish- 
able owing to " enchevetrement" of the layers. But his 
pupil Douliot (23) concluded that there was a single apical 
cell in all Gymnosperms, and a plero- periblem in most 


monocotyledons and some dicotyledons, while Koch takes 
the view that there is a generalised meristem without 
separate layers in Gymnosperms (21) and that only the 
dermatogen is separate in most Angiosperms (22). So 
that the "triple layer" theory of Hanstein and Van 
Tieghem is accepted by neither of these two most recent 
investigators as of general application, widely divergent as 
are their views inter se. Considering that the theory of 
the direction of ontogeny by the separation of different 
kinds of somatic idioplasm is now generally discredited, it 
is difficult to see what we gain by an adherence to the un- 
proved hypothesis of the strict separation of the initial 
layers, even if it is still a possible hypothesis. 

In the root apex on the contrary the plerome is in the 
great majority of cases sharply separated from the peri- 
blem, but even this rule is not universal. The sharp 
separation seems to be correlated both in root and stem 
with the formation of a slender compact cylinder. 

In Vascular Cryptogams, which nearly all possess either 
a single apical cell or a single layer of initial cells giving 
rise to the whole of the tissue of the axis, there is of course 
no question of a separation, at the apex itself, of initial 

The separation of the young cylinder behind the actual 
growing point is quite a distinct question from its separation 
at the apex. It is during the development of the cylinder 
that we get, usually at least, a distinct limit between it and 
the cortex which is often lost in the adult stem, and this is 
a point of great importance. 

Long before the stelar theory was originated, most of 
the great anatomists, who laid the foundations of our know- 
ledge of the histology of vascular plants, were practically 
agreed on the generality of this early separation. This is 
clearly shown in the terminology employed in designating 
the various regions. 

Thus Sanio (24), tracing from the apex the development 
of the various tissues, showed that in many cases the young 
pith first became separated from an outer zone, and that 
in the latter the "thickening ring" (really corresponding to 


Flot's "vascular meristem," i.e., the ring of tissue produc- 
ing the bundle system plus the "external conjunctive": 
shortly became differentiated from the peripheral zone or 
young cortex. In other cases {Euonymus and Berberis), 
the "thickening ring" appeared or began to appear before 
the young pith became separated from the " outer zone". 
Hanstein (25), as a consequence of his separation of the 
primary meristem into Dermatogen, Periblem and Plerome, 
makes the outer limit of the young cylinder, i.e., that 
between periblem and plerome, of primary rank. Russow's 
scheme (26), on the other hand, drawn from instances like 
those of Sanio's first group, 1 in which the young pith is the 
first tissue to become apparent, divides the young tissue 
produced by the general Protomeristem at the apex itself 
into Endistem (Sanio's young pith) and Existent (Sanio's 
" Aussenschicht "), the latter being separated into Mesistem 
(Sanio's "thickening ring ") and Peristem or young cortex. 
Thus the limit between "Mesistem" and "Peristem' is 
reduced to secondary rank. But De Bary (14, pp. 395-6) 
again sums up clearly in favour of the individuality of the 
plerome. 2 As a matter of fact the young pith often does 

1 Russow placed Hanstein's best instances, for example, stem of Hip- 
puris, and Roots, where there is a well-defined plerome at the apex itself, 
under the separate heading of "Axes with Combined Bundles ". 

2 The development of the pericycle is of great importance in this 
connexion. Sanio (24) showed in several cases that what we now call the 
pericycle was developed from the outer edge of the "thickening ring". 
Schmitz (27) confirmed this view in Berberis and Menispermum. Van Tieg- 
hem, however (5), based his conception of the pericycle entirely on the 
ground of adult comparative anatomy. This is explicitly stated (p. 152) in 
a remark he made at the close of a " Note sur le pericycle," read by 
D'Arbaumont (28) to the Botanical Society of France. D'Arbaumont had 
endeavoured to show that the sclerised portions of the pericycle, capping 
the phloems of the stem bundles in dicotyledons, were developed in 
common with the bundles themselves from the desmogen strands, and 
were thus often separate from the interfascicular pericycle. His account 
of the development of the continuous zone of fibres in Cucurbitacese and in 
Berberis is different, and indicates differences in the origin of the pericycle 
in various plants. It is unfortunate that no figures are given. Morot re- 
plied (29) that even if the pericycle, or parts of it, were developed dif- 
ferently in different plants, that made no difference to the validity or applica- 


become recognisable in comparatively bulky apices (owing 
to the early ceasing of longitudinal divisions, and the stretch- 
ing of its cells), before the outer limit of the young cylinder 
is defined. On the other hand, in the slender stems of 
many water plants, Hanstein's scheme applies with dia- 
grammatic precision, the outer limit of the cylinder being 
clearly marked at the apex, before there is any sign of a 
differentiation between pith and bundle ring. But these 
differences of precocity in the development of the various 
regions of the cylinder, depending, as they do, upon the 
subsequent duration and size relations of the regions are 
clearly of little importance to morphology. The important 
fact which remains is the clear separation, slightly sooner, 
or slightly later, of the young cylinder from the cortex, in 
at any rate the vast majority of cases. 

The separation thus made in development is, as a rule, 
more or less clearly maintained in the adult stem, though 
sometimes it is lost altogether. There is the possibility 
of a complete loss of a visible boundary between cylinder 
and cortex by the occurrence of irregular cell divisions in 
the young pericycle and inner cortex, together with a 
"shifting" (Verschiebung) of the original walls separating 
the two ; unfortunately we do not know if this takes place 
in some cases or not. But apart from such an occurrence 
the distinction between cylinder and cortex, once made, is 
always made, and the layer of cells which once abutted on 
the young cylinder is still the phloeoterma, not merely 
"theoretically," but in substance and in fact, however im- 
possible it may become to distinguish it from the surround- 
ing tissue. 

It is these facts which form the real developmental basis 
of the stelar theory. 

The phenomena (supposing them to be established) of 
real importance in the opposite sense, would be the occur- 
rence of stems in which the external limit of the cylinder is 
never clear, of stems, in a word, which never possess a 

tion of the term. The further pursuit of the theoretical implications of this 
statement would lead us into very deep waters, but it is clear that an ex- 
tended comparative investigation of the origin of the pericycle is needed. 


cylinder as such. While we could not admit that the stelar 
doctrine applied to such stems, we should probably be 
forced to the conclusion if their vascular system conformed in 
all other respects to the monostelic type, that the plants in 
question were derived from truly monostelic ancestors, whose 
descendants had lost the limit between cortex and cylinder. 

The Nymphseaceae, many of whose stems contain a 
large number of "scattered" bundles, seem to furnish us 
with examples of such plants. Caspary (27) states that 
the bundles are here developed in centripetal order : this 
would seem to indicate an analogy with those plants 
(Piperaceae, Begoniacese, etc.), which possess a proper 
bundle ring" and also younger bundles in the pith, rather 
than with the monocotyledonous type. In at least one 
member of the family, Victoria regia, which possesses 
a particularly large number of these "scattered" bundles, 
it appears that no well-defined cylinder is visible anywhere 
in the stem. 1 So here if anywhere we seem to have a real 
case of "astely". We cannot, however, say the same 
with certainty of any dicotyledonous stem with a single 
ring of bundles. Nageli's observations (28) indeed led him 
to the conclusion that the " cambial " strands were, as 
a rule, developed in the midst of a homogeneous ground 
tissue, but his conclusions, as we have seen, have been 
negatived by most subsequent observers. 

Turning to the vascular cryptogams we find that 
whether monostelic or polystelic, the stele or steles can be 
traced nearly up to the stem apex. The first formed peri- 
clinal walls do not indeed necessarily mark the limit of 
stelar tissue. They may cut off the pith, as in Equisetum 
or mark the middle of the cortex, as in many roots, or 
the outer limit of the ring of steles, as in many fern stems, 
or of the single cylinder, as in the stolon of Nephrodium 
(10, pp. 692 and 773-4). Clearly no special importance 
can be attached to these walls, and we certainly can- 
not use the fact that they mark off the pith in Equisetum, 

1 1 owe this information to the kindness of a friend in telling me the 
results of some unpublished observations. 


as Van Tieghem does, to support the view that the genus 
is really astelic. This argument depends on the assump- 
tion that these walls always separate stelar from extra- 
stelar tissue, which is not a fact, according to Van 
Tieghem himself (10, p. 774), and further, a similar line of 
reasoning would tend to show that the stems of a great 
many dicotyledons, namely, those in which the pith is the 
first tissue to be marked off, are also astelic. 


We have attempted in the foregoing pages to ex- 
hibit, as clearly as possible, the bearing of well ascer- 
tained facts of anatomy and development upon the stelar 
theory as developed by Van Tieghem and his pupils. We 
may appropriately conclude with an attempt to summarise 
the results to which we are thus led. 

We recognise in the central cylinder of the axes of the 
great majority of the higher plants an anatomical region of 
the first rank to be co-ordinated with the other great 
anatomical regions, the cortex and the epidermis. The 
central cylinder consists of vascular tissue (xylem and 
phloem) and conjunctive tissue (typically parenchyma). 
In the bulky typical * cylinder the vascular tissue is separ- 
able into distinct strands corresponding with its centres (or 
rather lines) of development, and giving to the cylinder a 
radial symmetry ; the conjunctive of such a cylinder is 
separable into distinct regions. Typically, also, the inner- 
most layer of cortex, which abuts on the cylinder is dis- 
tinguished by special characters. 

Reduced central cylinders are found in various stem 
structures, especially the thin stems of water plants. The 
reduction acts first on the conjunctive, which may (though 
rarely) quite disappear. This leads to the coalescence of 
the strands of vascular tissue into a more or less solid 
cylinder. Such a reduced cylinder is always sharply marked 
of! from the cortex. 

On the other hand we have stems in which it is im- 
possible to separate the conjunctive from the adjacent 

1 In Sach's sense of " most highly developed ". 


cortical tissue. When this is the case in the adult, it is still 
often possible to make the separation in the young stem. 

Naming the central cylinder a stele, we call all stems 
with a single cylinder monostelic. 

Stems in which we cannot make the separation in any 
part, and which are therefore not strictly monostelic, yet 
conform more or less to the monostelic structure in other 
respects, and are no doubt usually derived in descent from 
the monostelic type. 

Most Ferns and Selaginellas, and two genera of 
Phanerogams, while showing a monostelic structure in 
their hypocotyls, possess in their later formed stems more 
than one cylinder, each comparable in structure to the single 
stele of the hypocotyl. Such stems are known as polystelic. 
The steles of a polystelic stem may, however, take on the 
most various forms, and lose all the characters of the 
original cylinder ; several may even coalesce to form a 
structure indistinguishable from a single stele. As this, or 
indeed the converse case of a non-stelar vascular strand 
assuming the characters of a stele, may have happened in 
descent without leaving any traces of the transformation, we 
are not justified in asserting the homology of all steles or 
denying homology between steles and non-stelar vascular 
strands. Nevertheless the stele is undoubtedly a real and re- 
latively stable type in the arrangement of vascular tissue, and 
hence the name represents a real morphological conception. 

The vascular tissue of a leaf is arranged in one or more 
strands, each of which, bilaterally rather than radially 
symmetrical, is called a schistostele or meristele, representing, 
as it does, a part only of the stem cylinder. The meristele 
of a petiole may, however, simulate a stele. In most poly- 
stelic stems one or more of the stem steles directly enters 
the petiole, and the branches maintain more or less of the 
stelar character till near their endings in the lamina, where 
they become indistinguishable from collateral bundles. 

We are probably justified in supposing the monostelic 
type to be primitive in vascular plants, and we may assume 
the original stele to have been relatively simple. To the 
increase in bulk of the stem and correlated increasing de- 


mands for the supply of vascular tissue to leaves, the plant 
either responded by increasing the bulk of the stele and 
multiplying the number of its vascular strands, or by sub- 
stituting a number of simple steles for the original single 
one. This last occurrence happened once at least in the 
Pteridophyta (probably more often), and more than once 
among the Phanerogams. 

The primordial stele is represented at the present day 
by the single sharply defined stele of the embryo, which is 
maintained in the root and hypocotyl, and which passes over 
in the stem to one of the modern types of structure, 
necessary to the various demands of the leafy shoot. The 
arrangements at the apex of the latter are naturally adapted 
to form the particular type of structure in question, and 
can in no case be considered as representing an ancestral 


(1) Ph. van Tieghem. Recherches sur la symetrie de structure 

des plantes vasculaires. Introduction, pp. 5-29. La Racine, 
pp. 30-314. Annates des Sciences Naturelles, Botanique, 5 
ser., tome xiii., 1870-71. 

(2) Van Tieghem. Memoire sur les canaux secreteurs des 

plantes. Ann. Set. Nat. Bot., 5 ser., tome xvi., 1872. 

(3) FALKENBERG. VergleicJiende Untersuchnngen fiber d. Ban d. 

der Vegctationsorgane d. Monocotyledonen. Stuttgart. 1876. 

(4) MANGIN. Origine et Insertion des racines adventives. Ann. 

Set. Nat. Bot., 6 ser., tome, xvi., 1882. 

(5) VAN TIEGHEM. Sur quelques points de l'anatomie des Cucur- 

bitacees, p. 277. Bulletin de la Societe" Botanique de Fratice, 
tome xxix., 1882. 

(6) Morot. Recherches sur le pericycle. Ann. Sci. Nat. Bot., 1884. 

(7) Van Tieghem et Douliot. (a) Structure de la tige des 

Primeveres nouvelles du Yun-nan, p. 95. (b) Groupement 
des Primeveres d'apres la structure de leur tige, p. 126. (c) 
Sur les tiges a plusieurs cylindres centraux, p. 213. Bull. 
Soc. Bot. France, tome xxxiii., 1886. 

(8) Van Tieghem et Douliot. Sur la polystelie. Ann. Sci. 

Nat. Bot., 7 ser., tome iii., 1886. 

(9) Leclerc du Sablon. Recherches sur la formation de la tige 

des Fougeres. Ann. Sci. Nat. Bot., 7 ser,, tome xi., 1890. 

(10) Van Tieghem. Traite de Botanique, 2ieme edition, 1888-91. 


(il) FLOT. Recherches sur la zone perimedullaire. Ann. Sci. 
Nat. Bot., 7 ser., tome xviii., 1893. 

(12) VAN TlEGHEM. Remarques sur la structure de la tige des 

Preles. Journal de Botauiqne,tome iv., p. 365, November, 1 890. 

(13) Van TlEGHEM. Remarques sur la structure de la tige des Ophio- 

glossees. Journ. de Bot., tome iv., p. 405, December, 1890. 

(14) De Bary. Vergleichende Anatomie der Vegetationsorgane 

der Gefasspflanzen, 1877 (English edition, 1884). 

(15) STRASBURGER. Ueber den Bau und die Vorrichtungen der 

Leitungsbahnen in den Pflanzen. Histologische Beitrage iii., 

(16) VAN TlEGHEM. Pericycle et Peridesme. Journ. de Bot., 

tome iv., p. 433, December, 1890 

(17) Van TlEGHEM. Sur la structure primaire et les affinites des 

Pins. Journ. de Bot., tome v., p. 265, etc., August, 1891. 

(18) PAUL Zenetti. Das Leitungssystem im Stamm von 

Osmunda regalis L. und dessen Uebergang in den Blattstiel. 
Botanische Zeitung, April, 1895. 

(19) Poikault. Recherches anatomiques sur les Cryptogames 

vasculaires. Ann. Sci. Nat. Bot., 7 ser., tome xviii., 1893. 

(20) D. H. Scott. Recent work on the Morphology of Tissues 

in the Higher Plants. "Science PROGRESS," vol. i., 
August, 1894. 

(21) L. KOCH. Ueber Bau und Wachsthum der Sprossspitze der 

Phanerogamen. Pringsheinis Jahrbiicher f. zvissenschaftliche 
Botanik, Bd. xxii., 1891. 

(22) L.KOCH. Die vegetative Verzweigung der hoheren Gewachse. 

Pr.J., Bd. xxv., 1893. 

(23) DOULIOT. Recherches sur la croissance terminale de la tige 

des Phanerogames. Ann. Sci. Nat. Bot., 7 ser., tome xi., 1890. 

(24) San IO. Vergleichende Untersuchungen iiber die Zusam- 

mensetzung des Holzkorpers. Bot. Zeit., 1863. 

(25) Hanstein. Die Scheitelzellgruppe, 1868. 

(26) Russow. Vergleichende Untersuchungen betreffend die 

Histologic . . . der Leitbiindel-Kryptogamen, u. s. w. 
Memoires de V academic imperiale des Sciences de St. Pcters- 
bourg i 7 ser., tome xix., 1872. 

(27) Schmitz. Ueber die Entwicklung d. Sprossspitze d. Phanero- 

gamen. Halle. 1874. 

(28) D'Arbaumont. Note sur le pericycle. Bull. Soc. Bot. de 

France, tome xxxiii., p. 141, 1886. 

(29) Morot. Reponse a la note de M. D'Arbaumont sur le 

pericycle, ibid., p. 203. 

(30) Nageli. Beitrage zur tvissenschaftliche Botanik, i., 1858. 

A. G. Tansley. 



SINCE I have shown that protoplasm in the simplest 
form in which it is known to us may not be regarded 
as having an organisation in the sense in which that term 
has any meaning, and since it is a waste of time to discuss 
the use of the term when it has no meaning, we may more 
profitably turn to the question whether protoplasm has a 
structure, and if so, what kind of structure? Is it essenti- 
ally the same in all the kinds of protoplasm which have 
been studied, and is it of the same kind as the structure of 
tissues and organs of metazoa or is it of a different kind ? 
For it must be insisted upon that one may deny to proto- 
plasm an organisation, in the proper sense of the term, and 
yet one may consistently attribute to it a structure, even a 
very complex structure. But that structure need not be 
called an organisation, to do so is to confuse two clear 
issues. It is worth while to emphasise this point, for some 
people think it very inconsistent to affirm that protoplasm 
has a complex structure and at the same time to deny that 
it is organised. 

I conceive that the view that protoplasm is composed of 
granules, which are either biophors or secondary aggregates 
of biophors, has been sufficiently refuted by Butschli's re- 
searches on hyaline protoplasm already referred to. The 
hyaline pseudopodia of Gromia show no trace of granules, 
not because the granules are too small to be seen, for the 
highest powers of the microscope reveal in the protoplasm, 
at the moment of its protrusion to form a pseudopodium, a 
structure which is not granular, namely, an alveolar structure, 
and if granules were present they must necessarily be sought 
for in the alveoli or in the alveolar walls. But they are to 
be found in neither, so it may be affirmed that in the 
simplest form of protoplasm there are no granules, a 

circumstance which deprives the theory of biophors of much 



of its weight. Of course it may be objected that the 
alveolar walls and contents may be composed of biophors 
so small as to defy detection ; such an objection must be 
defended on theoretical grounds, and I will deal with it 
presently ; just now I will confine myself to the considera- 
tion of the visible structure of protoplasm. 

After rejecting the granular theory we have a choice of 
several others ; the fibrillar theory, the reticular theory, and 
the alveolar theory of Biitschli. It would take too long for 
me to examine these several theories in detail ; it has 
already been done by Biitschli (loc. ciL, p. 177), and still 
more recently by Yves Delage, 1 if I were to undertake the 
task I should only give a resume of their arguments. 
For my own part I am strongly inclined in favour of Biit- 
schli's " Wabenlehre ". 

For some reason or other Biitschli's account of the 
structure of protoplasm has not, to use a common ex- 
pression, " caught on ". Possibly because it was published 
at a time when men's minds were occupied with the more 
alluring prospect offered by the granular theory of proto- 
plasm, with all its delusive hopes of an explanation by means 
of biophors, and primary organisation of the phenomena 
of heredity, and of all the vital processes. Possibly also 
because Biitschli himself pushed the analogy between micro- 
scopic foams and protoplasmic structure too far. But if 
his theoretical considerations are put aside, there is a great 
deal to be said for his fundamental views. The alveolar 
structure which he describes may be demonstrated in many 
various forms of protoplasm. It is particularly obvious in 
Pelomyxa, in which form the larger vacuoles serve admir- 
ably as a contrast between the finer alveolar structure which 
he claims to be common to all protoplasm and the grosser 
vacuolar structure which is often mistaken for it. I have 
myself identified the alveolar structure in a considerable 
variety of protozoa, and in a number of tissue cells, and I 
have succeeded in making Biitschli's artificial amoebae, and am 

1 Yves Delage, La Structure du Protoplasma et les Theories sur 
V Heredite et les grands problems de la Biologie generate. Paris : C. 
Reinevald et Cie, 1895. 


convinced of the close analogy in structure between the 
artifact and the natural product. The resemblance between 
the two is exact, and it is astonishing. The optical char- 
acters of the artificial product are explained, on physical 
grounds, as the outcome of a certain structure, namely, an 
alveolar structure. The identical optical characters of pro- 
toplasm may surely be explained on the same grounds. It 
is not pushing analogy too far to say that identical optical 
characters are the result of identity of structure. The 
analogy is somewhat strained when it is sought to prove 
that the identity of the streaming movements in the arti- 
ficial product with those in protoplasm are attributable to 
the same physical causes. The chemical constitution of 
the. two bodies is so different that the phenomena observed 
might be regarded as secondary. Nor is the identity 
absolute, for Biitschli himself points out that the induced 
currents in the surrounding medium take place in the re- 
verse sense in an amoeba to what they do in the case of 
the microscopic foam. I cannot think that the criticism of O. 
Hertwig invalidates Biitschli's theory seriously. Hertwig 
says that lamellae of oil consist of a fluid which is not 
miscible with water. If the comparison between the 
structure of an emulsion and the structure of protoplasm 
depends on something more than a superficial resemblance, 
then the lamellae of plasma which are compared with the 
lamellae of oil must consist of a solution of albumen 
or of a fluid albumen. But a solution of albumen is 
miscible with water, and therefore it would mix with the 
contents of the alveoli : emulsions of albumen must be formed 
with air, not with water. To this Biitschli answered that 
the framework of plasma consists of a fluid composed of a 
combination of an albumen and a fatty acid, which was 
therefore not miscible in water. Another obvious answer 
is that living plasma is not a simple albuminous solution, 
for if it were most protozoa could not exist, they would 
immediately dissolve in the water in which they live. 
Whether a fatty acid exists in combination with the plasma 
or not, there is something in the constitution of living- 
plasma which differentiates it from albumen, for it does not 


dissolve in water ; dead plasma on the other hand becomes 
albumen and dissolves speedily. What that something is 
I do not venture to suggest ; could we ascertain what it is, 
no doubt we should have discovered the solution to the 
riddle of life. Hertwig says that the structural elements of 
protoplasm, be they filaments, or reticular, or lamellae, or 
alveoli, or granules, or what else, have a fixed state of 
aggregation. Protoplasm is no mixture of two immiscible 
substances such as water and oil, but consists of a union of 
fixed organic material particles with abundant water. This 
is but a verbal statement of the facts and is no explanation, 
but he adopts later on {Joe. tit., p. 49) Nageli's micellar 
theory as an explanation. No doubt it is the best explana- 
tion possible, but it again does not give more than a verbal 
explanation of the remarkable and fundamental phenomenon 
that protoplasm, be its structure what it may, does not when 
alive dissolve in water, but when dead it becomes some- 
thing else which readily dissolves, provided of course that it 
is not killed by means which coagulate the albumens into 
which it is converted at death. 

I shall recur again to the micellar theory, for the pre- 
sent purpose it is sufficient to say that it is not inconsistent 
with Biitschli's " Wabenlehre," 1 and might even be pressed 
into service to explain why the plasma does not mix with 
the watery alveolar contents without the necessity of calling 
fatty acids to aid. 

Supported by these considerations, and by a considerable 
mass of objective evidence, I venture to think that Btitschli 

1 Biitschli criticises the micellar theory and the analogous theory of 
"inotagmas" put forth by Engelmann. He does not accept either, but 
does not give in their place any theory of the ultimate compositions of the 
substances which form the alveolar framework and contents, except that 
(p. 309) he says, " a series of reflections . . . led me to suppose . . . that 
the chemical basis of the framework substance must be formed by a body 
which has arisen from a combination of albuminoid and fatty acid mole- 
cules." Such a combination must mean the formation of a chemical unit 
of a higher order than the molecules which enter into its composition, and 
for my purposes such a chemical unit is a micella. In this limited sense 
the acceptance of a micellar structure is not incongruous with the " Wa- 
benlehre ". 


has given a true account of the minute structure of proto- 
plasm, so far as it can at present be determined by optical 
means. And I even venture to prophecy that when the 
history of the biological work of this half century comes to 
be written some half century hence, the theories of biophors 
and plasomes and the such like will have merely a historical 
interest, whilst the work of Biitschli will be regarded as the 
most sagacious and far-sighted contribution of our time to 
this momentous question. In saying this I do not wish 
to declare my adhesion to the more theoretical part of 
Biitschlis work, but only to his account of the microscopic 
structure of protoplasm. 

Even if one were to accept his explanation of the 
streaming movements there would remain all the other 
phenomena of life to be accounted for, and they are inex- 
plicable on the visible structure of protoplasm, even if it be 
an alveolar structure. 

Underlying the visible structure then there must be an 
invisible structure, which is the cause of the phenomena. 
This admission once made, the claims of the rival theories 
of biophors, plasomes, plastidules and what not, again press 
themselves on our attention. Now it is to be remarked 
that the most cautious and thoughtful theorists do not claim 
that their hypothetical units are an explanation of life. 
Weismann categorically denies that his theory of the germ 
plasm is a theory of life, it is only a theory of heredity, but 
he goes so far as to suggest that a workable explanation of 
the more complicated vital phenomena may be the surest 
indication of the path which will lead to an explanation of 
the more simple (loc. ciL, p. 21). 

Others, however, are not so cautious, and in any case 
there is this feature common to all, that they aver on the 
one hand that vital processes are so complicated that they 
cannot be explained by a physico-chemical theory of the 
constitution of protoplasm, and that therefore we must 
assume the existence of ultimate vital units or biophors : 
on the other hand, after endowing these biophors with all 
the attributes of life, they say that they have a comparatively 
simple molecular constitution upon which the phenomena 


which they exhibit depend. In fact they describe essenti- 
ally similar functions in biophors and in cells, but they 
allow a physico-chemical explanation in one case and 
disallow it in the other. This contradiction has been 
noticed by others, and it has never been satisfactorily 
explained away. Whitman draws attention to it, and 
observes that no one, as far as he knows, has looked upon 
the unit as anything more than the seat of the mystery. 
This is true, but it is no reason for putting" the mystery in 
a small bag instead of a big one. He defends the theories 
of smaller units, however, by saying that they have ex- 
tended our knowledge of organic mechanism {Joe. cit., 
prefatory note, p. vi.). This again I believe to be true, 
but not quite in the sense in which Whitman apparently 
means it to be. The theories of minute independent vital 
units have, I believe, led many on the wrong track as 
regards vital mechanism ; the attacks on such theories are 
leading to a considerable extension of our knowledge in 
this direction. The ultimate vital units confessedly do not 
remove the mystery ; ultimately the explanation of life 
must be a chemico-physical one ; there is no alternative 
but a vitalistic theory, and this is not admissible in science. 
The strongest ground, viz., the granular hypothesis, for 
assuming the presence of vital units is removed by the 
observed constitution of hyaline protoplasm, and finally 
none of the assumed aggregates of units which are admitted 
to be visible, are identified with various sorts of granules 
and considered to constitute units of a higher order, have 
ever been shown to be capable of leading an independent 

On the other hand there is a oeneral consensus of 
opinion that protoplasm is not a simple organic compound. 
Its unit is not the molecule, but an aggregate of molecules 
forming a unit of a higher order to which the molecule 
stands in the same relation as the atom does to the mole- 
cule. It is also admitted that these molecular aggregates 
may exist in many different kinds in protoplasm. Such a 
conception is absolutely necessary for the explanation of 
the most simple properties of organic bodies, for example, 


their optical properties and the imbibition of water. But it 
is a physico-chemical conception, and the molecular aggre- 
gate need not and should not be endowed with independent 
vital powers. Such a molecular aggregate is the micella. 
In accepting the micella one may attribute any amount of 
complexity to protoplasmic structure without for a moment 
admitting that it is a cono-eries of elementary organisms. 
Nor need we admit all the theories which Nageli has tried 
to establish as the necessary consequences of the assumption 
that there are such things as combinations of polyatomic 
molecules into groups of a higher order. As I have already 
said, it was pointed out by von Sachs that even in the 
region of pure chemistry it is necessary to assume that polya- 
tomic molecules are grouped into closer molecular unions, 
thus giving rise to chemical properties which did not belong- 
to the individual molecules. But in the region of pure 
chemistry such a grouping is not called an organisation, 
nor is there any reason why it should be called an organisa- 
tion in the present case. Let us be perfectly definite and 
say that by a micella we mean a combination of polyatomic 
molecules into closer union to form a group ; nothing more, 
except in so far as we may reason on chemico-physical 
grounds as to the behaviour of such groups and their 
relations inter se. For instance (I am quoting from O. 
Hertwig's summary of this part of the micellar theory) : 
"The micellae exert an attraction both on water and on one 
another, whereby the phenomena of swelling may be ex- 
explained. In a dry organic body the micellae lie close to 
one another, separated only by exiguous envelopes of 
water : these latter enlarge considerably during imbibition, 
since the attractive forces between the micellae and water 
are at first greater than between the micellae themselves. 
The micellae are separated from one another by the imbibed 
water as it were by a wedge ; but an organised body does not 
arrive at a condition of solution, since the attraction of the 
micellae for water diminishes in the course of their separa- 
tion from one another, at a greater rate than the attraction 
of the micellae for one another, and therefore, when the 
watery envelopes have attained a certain size, a condition 


of equilibrium, the limit of imbibition is reached." And 
also: "Since particles of water may be held fast on the 
surfaces of the micellae by molecular attraction, so also 
other matters (lime and siliceous salts, colouring matters, 
gelatin compounds, etc.) may be deposited on them after 
they have been taken into the organic body in a state of 
solution ". So far as my physical knowledge enables me to 
form a judgment, attributes such as these may justifiably be 
ascribed to micellse on purely physical grounds and their 
importance can hardly be overestimated, since the last 
passage quoted affords a hint as to the nature of the essen- 
tially vital process of assimilation. It is not my business 
now to develop a complete theory ; I doubt indeed whether 
a complete theory is possible in the present state of our 
knowledge. I have done sufficient for present purposes if 
I have succeeded in indicating what ideas we may justifiably 
hold on the subject of protoplasmic structure, and I believe 
that I have given some good grounds for justification of the 
views that ; (i) the ultimate visible structure of protoplasm 
is an alveolar structure ; (2) that the invisible structure of 
protoplasm is a "micellar" structure in the sense defined 

But before I proceed I must enter a caveat against 
being considered as an adherent of the micellar theory of 
Nageli. I cannot enter here into my reasons, but I may 
say that the further theories which Nageli assumes to 
be the necessary consequences of the existence of micellae, 
do not appear to me to be necessary consequences at all ; 
indeed I part company with him at once when I express my 
conviction that the hypothesis of a micellar structure is 
compatible with the alveolar structure described by 
Butschli. 1 

1 Since the above argument was first written out the work of Yves 
Delage has come into my hands. It is most gratifying to find that the 
opinions of so distinguished an author accord so exactly with my own. The 
reader who finds my argument involved and laborious may turn with profit 
to Delage's book, in which he will find a lucidity of expression and a 
precision in argument which I can only envy without hoping to imitate. It 
is worth while quoting the following passages here: "On peut accorder 


I may now anticipate the objection which is certain to 
be raised that the visible and invisible structure which I 
assign to protoplasm is utterly inadequate to explain the 
phenomena of life. It is inadequate and it is intended to 
be inadequate. Were I to pretend that it is adequate I 
should be running counter to all the lessons taught by our 
experience of living things. The structure which I have 
assigned to protoplasm applies particularly to that simplest 
known form of it which we rarely meet with, but which we 
do meet with in exceptional cases, for instance in the pseudo- 
podia of Gromia dujardini. But separate a protoplasmic 
corpuscle formed by the thickenings of the thread-like pseu- 
dopodia of this species from the rest of the animal ; the cor- 
puscle separated is not any longer capable of an indepen- 
dent existence, it soon perishes, it has all the structure 
which I have described, but it is not capable of in- 
dependent life. Clearly then life is not the outcome of this 
structure, though the structure may play its part, and no 
unimportant part in the life processes. 

When I have been speaking of protoplasm I have 
obviously been confining my attention to that form of it 
which is now generally distinguished under the name of 
Cytoplasm. Cytoplasm taken by itself is not living matter 
in the sense that it is capable by itself of maintaining an 
independent existence. The experiments of Nussbaum, 1 of 
A. Gruber and Verworn, confirmed by other observers, have 

a l'auteur (Nageli) ses Micelles. Leur constitution, leurs proprietes n'ont 
rien que de tres admissible. Bien que leur mode de generation ne soit 
guere probable, il n'y a aucune raison positive pour le repousser. Mais 
l'arrangement des micelles et la structure de l'idioplasma sont invraisem- 
blables au plus haut point. Nous avons demontre, au cours de notre 
expose, que cet arrangement n'est pas de tout, com me l'auteur l'avance, le 
resultat necessaire du seul jeu des forces moleculaires initiates ce n'est 
qu'a grand renfort d' hypotheses etagees l'un sur les autres qu'il arrive a 
faire disposer les Micelles en Files, les Files en Faisceaux, les Faisceaux en 
Cordons et les Cordons en un Reseau repandu dans tout l'organisme." 

1 It was Nussbaum who first introduced the method of dividing in- 
fusoria by artificial means, and the credit of having devised this very useful 
class of experiment belongs to him. In my previous article I inadvertently 
assign it to Gruber. 


shown that pieces of cytoplasm cut off from the remainder 
of a protozoon are incapable of maintaining life and soon 
perish. If, on the other hand, a fragment of cytoplasm 
similarly cut off contains nuclear matter, it is shown to con- 
tain the attributes necessary to life, for the fragment does 
not perish but reconstitutes itself and becomes an inde- 
dependent living being. The converse also holds good. 
A nucleus or a fragment of a nucleus isolated from a 
protozoon, is incapable of life and perishes. But a nucleus 
or a fragment of a nucleus in conjunction with a fragment 
of cytoplasm is capable of life and constitutes an indepen- 
dent living being. The reasonable inference is that cyto- 
plasm plus nuclear matter is indispensable for the per- 
formance of vital functions. 

Now cytoplasm plus nuclear matter constitutes a cell. 

I have elsewhere discussed at some length the definition 
of a cell, 1 and I have defined it as a corpuscle of protoplasm 
which contains nuclein. In the present state of our know- 
ledge this definition seems the only one possible. The cell 
then consists of two essential substances, cytoplasm and a 
substance which is different from cytoplasm, both structurally 
and in chemical constitution, namely, nuclein. In a great 
majority of cells other substances are present which are 
different from both of these. Such substances are the 
centrosomes, that modification of cytoplasm which is called 
archoplasm, amylum and aleurone grains and so forth. As 
far as we know, however, these substances are not essential 
to life, but are secondary products characteristic of dif- 
ferentiated cells. Recent researches on the structure of 
Bacteria and Oscillaria justify the assertion that cells exist 
in which these substances are absent. We know next to 
nothing about the presence or absence of centrosomes and 
archoplasm in the Protozoa, and it may be that further 
investigation will lead us to the conviction that these two 
are as essential to the life of these forms as the presence of 
cytoplasm and nuclein. Maybe not ; in any case it does 

1 Quarterly Journal of Microscopical Science, vol. xxxviii., p. 137, 


not matter for present purposes. It is sufficient to know 
that two substances, cytoplasm and nuclein, must be brought 
together or life cannot exist, and that it does exist in 
organisms in which these substances, and these only, can be 
detected, viz., in Bacteria. This statement may appear some- 
what hazardous, seeing that the presence of a nucleus is 
denied in several living beings, in bacteria, for instance, and 
in yeast. A nucleus in the sense of a centralised body is 
certainly absent in these and in many other forms, but 
Biitschli has demonstrated the presence of nuclein in 
Oscillaria in Bacterium lineola. As for Saccharomyces it 
undoubtedly contains nuclein, for Raum has prepared it 
from yeast cells, and the most recent observer, Macallum, 1 
is of the opinion that the nuclein is distributed through the 
cytoplasm but also aggregated in the so-called granules of 

The statement therefore can scarcely be called hazardous, 
and it is really warranted by the facts at our disposal, for 
the more carefully that researches are made, and the more 
delicate the methods of investigations employed, the more 
is the presence of nuclein demonstrated where it was not 
previously supposed to exist. 

Macallum's paper, by the way, is of great interest, for he 
shows that nuclein is essentially the iron-holding substance 
in cells. Knowing as we do the close connection there is 
between the presence of iron and the due performance of 
the vital processes, this observation opens up a fruitful 
source of inquiry as to the dependence of life on chemical 

Throughout this argument I have tried to stick to the 
rule of drawing legitimate inferences from observed facts 
without wandering into the obscure regions of hypothesis. 
If I have been successful and have fairly stated the facts, 
and have drawn legitimate inferences, the conclusion which 
I come to must be admitted to be of considerable weight. 

1 A. B. Macallum, "On the distribution of Assimilated Iron Com- 
pounds, other than Haemoglobin and Haematins, in Animal and Vegetable 
Cells," Quart. Jour. Mir. Sri., vol. xxxviii., pp. 175-274, 1895. 


The conclusion is this : that life is possible only when two 
(or more) substances of complex chemical constitution are 
brought together, and that when these two (or more) substances 
ai'e brought together we have before us a cell. The cell there- 
fore is the vital unit /car' e^o^V. The component parts of 
the cell are not vital units, for by themselves they are in- 
capable of life; they are the auxiliaries, the indispensable 
auxiliaries of life, but they are not themselves living. 

This is not a theory of life, and it does not pretend to 
be one. It is the generalisation which the facts seem to 
warrant, and if it be true, as I believe it must be true, it is 
entirely inconsistent with the whole group of theories based 
upon hypothetical biophors, gemmules, plasomes, physio- 
logical units, plastidules et hoc genus omne. Those theories 
are false. And the cell theory is not inadequate, but it is 
the only theory which our knowledge of structure and of 
life processes permits us to adopt, at least if we confine 
ourselves to that part of it which is essential, namely, that 
there is one general principle for the formation all tissues, 
animal and vegetable, and that principle is the formation of 

Cells are the ultimate vital units, though they are not 
the ultimate structural units ; they are the Lebenstrager, or 
biophors, and there are no living individuals lower than 

As I have made an effort to stick to facts and have 
slighted hypotheses, I shall doubtless incur the profound 
contempt of those superior persons who find no mental 
repose in things which can be clearly apprehended, but 
must leave the material support of earth and seek for rest on 
the unsubstantial pillows of cloudland. They will have 
abundant scope for exercising their contempt, for my con- 
clusion explains nothing, and gives no clue to the problems 
of heredity. 

As I have said in the earlier part of this essay, I have 
no intention to discuss here the complicated problems which 
are involved in the question of heredity. I take my stand 
on the position from which I started, namely, that if minute 


vital elements occur at all, those same elements which make 
life possible and control assimilation and growth must also 
be the agents in bringing about the phenomena of heredity. 
I have shown that minute vital elements smaller than 
cells cannot be believed to exist, and it is clear that the 
phenomena of heredity cannot be explained by things 
which have no existence. This is a sufficient answer to 
those who would say that the phenomena of heredity are 
such that we must make use of a hypothesis of minute 
vital elements, which are at once the bearers of the vital 
qualities and the bearers of the heritable qualities (the his- 
toric properties if the expression is preferred) of protoplasm. 
It is not true that a theory of heredity is impossible unless 
such elements are postulated. Delage has brought forward 
a theory of heredity which discards altogether the use of 
hypothetical biophors. I pass no criticism on his theory, 
favourable or unfavourable, but call attention to it merely 
for the purpose of showing that a theory without biophors 
is possible. It is no argument to say that the theories 
based on ultimate vital units have largely extended our 
knowledge of heredity. The Ptolemaic system of astronomy 
largely extended men's knowledge of the movements of 
the heavenly bodies, but it was not on that account a true 

Moreover, it will be hardly fair to twit me with the 
fact that I renounce, for the present, an attempt to explain 
the most complicated manifestations of life, for this is only 
an essay, and makes no pretence to be the development of 
a doctrine. 

It is not my present intention to frame hypotheses, not 
because I undervalue the use of hypothesis, but because I 
regard the first necessary step to be the formation of ideas 
appropriate to the facts. 

Dr. Whitman has recently written quite a nice little 
lecture on the subject of fact and theory, and has directed 
it against myself in particular, winding up with a trenchant 
paragraph to the effect that the claim to a monopoly of 
fact reflects an arrogance which seems to be epidemic. 
This homily is fortified by quotations from von Baer, 


Goethe, Huxley and Whewell. Now I never claimed 
a monopoly of fact, but that facts should receive a due 
share of recognition. Mutual service, as Whitman says, 
is the principle which ties theory and fact together ; quite 
so, but when theory runs altogether away from fact, the 
mutual service is wanting. Fact is a slow servitor, and 
drags heavily on the impatient feet of theory. The quota- 
tions from Goethe and Huxley do not lend support to the 
practice of making hypotheses, rather the contrary. "Ex- 
perience. Reflection, Inference " is an excellent motto, but 
inference does not mean making hypotheses, nor yet does 
the necessary process of generalisation and classification 
which Huxley recommends. The passage quoted from the 
last-named author condemns the mere cataloguing of facts 
under the name of Science, but it does not countenance the 
reckless use of theory. As for Whewell's aphorism, let me 
commend to Whitman a study of what that author says 
with regard to the failure of the Greek schools of philo- 
sophy. They did not fail, he says, because they neglected 
facts ; the Aristotelian school may be held to have surpassed 
the moderns in its appreciation of the value of facts. The 
Greeks certainly did not fail for want of boldness in theor- 
ising, nor for want of acuteness, of ingenuity and power of 
close and distinct reasoning. Nevertheless with all help 
from the twin-service of fact and theory their philosophy 
was a failure, and why ? Because, as Whewell points out, 
their ideas were not distinct and appropriate to the facts. 
May not the same thing be said of many of the theories of 
cell life and of heredity which have been so much in vogue 
in the last few years ? It was my object when I wrote on 
Epigenesis and Evolution to show that some ideas then 
current, were not appropriate to the facts ; it has been my 
object in the present essay to show that certain theories on 
cell life, beautifully constructed and ingeniously defended as 
they have been, are not appropriate to the facts. I am far from 
undervaluing the use of theory, and when I took occasion 
before, as I have done again now, to emphasise the impor- 
tance of attention to fact, I was not quite so ignorant nor 
so arrogant as Whitman supposed. The motto of Goethe 


might well have been reversed for the adornment of the 
title pages of some works of the last twenty years. " Theory, 
reflection, experience," the last named to be fitted in as best 
it might. 

Since the above passages were first written the great 
work of Yves Delage has came into my hands. Mine is 
not the only voice crying out in the great wilderness of 
theories. This new voice, however, is far greater and 
more powerful than mine. The reader who may be uncon- 
vinced by my clumsy argumentation should turn to the 
pages of Delage. For clear and candid exposition, trenchant 
criticism, and rigorous exposure of defects of reasoning, 
they are unsurpassed. Now that this part of my work is 
ended I feel that it need never have been begun, for all 
that I have had to say has been said in greater detail and 
with much greater force by Delage. 


THE normal life cycle of ferns, owing to the micro- 
scopic character of their reproductive apparatus, 
long baffled the comprehension of botanists. But some 
half a century ago, starting with the observations of Naegeli 
and Suminski and culminating in those of Hofmeister, 
the whole course of their ontogeny has been cleared up. 
The fern plant, as ordinarily so-called, produces on the 
back of its leaves or fronds, countless numbers of spores, 
which are formed within minute capsules or sporangia. 
When these spores germinate they give rise, not to a new 
fern plant, but to a leaf-like scale — the Prothallus. Upon 
the lower surface of this the sexual organs arise, and within 
them the sexual cells themselves are differentiated, and as the 
result of the fertilisation of one of the female cells or 
oospheres, by the male cell or antherozoid, a new fern plant 
arises. Thus in normal cases a regular alternation of a 
sexual with a sexless generation is seen. But although 
this is the course followed by the vast majority of the ferns 
which have been hitherto investigated, it is not the only 
one open to the plants. Thus Prof. Farlow in 1874 dis- 
covered that the formation of the sporophore (fern plant) 
generation might arise directly from the oophore (prothallus) 
generation, without the intervention of sexual organs, by a 
process resembling ordinary budding. De Bary, who 
followed the matter further, found that several ferns other 
than that examined by Farlow reproduced themselves in 
the same fashion, to which phenomenon the name of 
Apogamy was given, the marriage link being eliminated. 
Curiously enough De Bary found that a variety of one of 
our most vigorous British ferns reproduced itself constantly 
in this asexual manner, though the common form exhibited 
no abnormality in this respect. Recently, however, L. 
Kny, 1 pursuing these investigations further, has found the 

1 Entivickehing von Aspidium Filix mas. Sk'., i Theil., L. Kny, 


normal form to reproduce itself in both ways, and since his 
asexual examples occurred in thickly-sown pots, it would 
appear to be due to some extent to a starved condition 
induced by overcrowding, which checks the formation of 
the archegonia, and leads to the simple budding in their 
place. In all these instances the young plants are en- 
gendered upon precisely the same spots on the prothallus 
as the sexual one would occupy, and as their development 
and appearance are identical, it is only by preliminary 
watching that their apogamic origin can be determined. 

A case of Apogamy (or rather two cases), however, 
recently occurred in a sowing of my own, which is quite 
distinct from any I have seen described. A sowing of a 
plumose variety of Athyrium jilix foemina failed almost 
entirely, only two or three prothalli surviving. One of 
these after growing very large, nearly half an inch across, 
remained perfectly dormant the whole of the summer ; 
early in the autumn, however, the edge of the prothallus 
began to grow out and upwards in two places, and eventu- 
ally two slightly curved horns, 1 each about one quarter 
inch long, developed perpendicularly, one on each side of 
the indentation or sinus common to most prothalli. Later 
on, at a short distance from each tip, a small whitish bulbil 
appeared and these increased in size until the circination of 
several fronds was plainly visible, a small crown or caudex 
being developed. No roots, however, were emitted, and 
the two little plants, both, be it remarked, identically 
situated and very like in form, were evidently supported by 
the prothallic root-hairs, though by this time most of the 
prothallus was brown and dead. Subsequently I placed a 
piece of loam in contact, and into this both plants rooted 
and fronds were sent up, the first of which had no less than 
ten pinnate divisions on either side. It was thus, it will be 
seen, very different from the usually simple primary fronds 
produced either sexually or apogamously heretofore. Later 
on still, noticing that the tips of the horns were showing 
signs of dilating, I cut these off with a razor and laid them 

1 Gard. Chronicle, 10th Nov., 1894. 


down, two apparently normal and full-sized prothalli being 
the present result. In this case it will be noted that both 
plants were far removed from the usual site of reproduction, 
and both in this respect and in their vigorous development 
are differentiated from previously cited cases of apogamy. 
The second case alluded to occurred on another prothallus 
in the same pan, wherein the bulbil developed likewise 
upon a horn-like excrescence, but on the centre of the upper 
surface of the prothallus. This bulbil has developed into 
what is so far a very weakly plant of a different type to the 
others, but otherwise presenting no special feature. 

Until 1884 the Prothallus had always been regarded as 
necessarily the offspring of the spore, but in the autumn of 
1883 a presumed barren variety of Athyrium jilix fcemina 
{var. C/arisstma) was sent me for examination. For 
twenty years this plant had been observed to produce an 
immense number of apparent sori, but no spores were 
ripened or shed, and no offspring had consequently been 
raised. Some previous observations on dorsal bulbils, i.e., 
bulbils associated with the spore heaps in this same 
species, led me to the opinion that these apparent sori, 
which consisted of green pear-shaped masses instead of the 
capsules proper to spores, did not represent bulbils, but 
some abnormality in the development of the sporangia. 
To test this I laid down portions of the fronds, and 
to my intense surprise these pearshaped bodies com- 
menced at once to grow into prothalli, their tips dilating 
and spreading, while root-hairs and subsequently both 
archegonia and antheridia appeared in abundance. I at 
once gave a note of my observations at the Linnean 
Society x as demonstrating the development of the prothallus 
without the agency of the spore. The following season, 
pursuing my culture, I was able to exhibit a number of 
plants and such material as satisfied the society of the 
facts put forward. 2 Prof. F. O. Bower 3 then undertook 

1 " Observations on a Singular Mode of Development in the Lady Fern 
{Athyrium filix fosmina)" Linn. Soc. Journal Botany, vol xxi., p. 354-7. 

2 " Further notes on ditto,'' ibid., vol. xxi., pp. 358-60. 

3 " On Apospory in Ferns (with special reference to Mr. Charles T. 
Druery's observations),'' F. O. Bower, ibid., vol. xxi., pp. 360-68. 


the further investigation of the case, and found that the 
development of the sorus or spore heap went as far as the 
formation of the stalk of the sporangium or spore capsule, 
but at that stage it stopped and a vegetative growth set in 
to form the clusters of pear or club-shaped bodies which 
eventually went through the normal evolution of prothalli 
and sexual plants. Mr. G. B. Wollaston followed by 
providing material from a variety of Polystichum angulare 
in his possession, wherein the elimination of the spore and 
the entire soral apparatus was so complete that the prothalli 
were developed from the slender-pointed tips of the ultimate 
divisions of the fern-frond. Padley, P. ang. var pule her rinmm 
was the plant in question, and as it chanced that several 
other varieties of the same type existed, though found at 
widely sundered spots in England, it resulted that Dr. 
F. W. Stansfield and myself found the same character in 
two of them. Prof. Bower further observed that soral 
apospory, i.e., the form first noted, was also present on 
Padley's plants, and this too we, Dr. Stansfield and my- 
self, confirmed in the others. We have in these four 
examples, and in the genus Polystichum especially, ample 
proof that the spore is not an essential preliminary to 
the existence of the Prothallus, but that the latter may 
be developed direct from the tissues of the Sporophore, 
precisely as this latter in Apogamy may be developed 
direct from those of the oophore. 1 Curiously enough 
the next case which came before the writer's notice 
was an aposporous seedling of the same variety of Lastrea 
(Aspidium) determined by De Bary as being persistently 
apogamous, viz., Lastrea pseudo ?nas var. cristata. This 
case was distinct from previous ones as it was a young 
plant and not an adult, which produced the prothalli. The 
tip of the second frond evolved from the prothallus (the 
first was eaten off and was not seen) bore a prothallus of 
the normal form. Subsequently this and the succeeding 

1 Professor F. O. Bower subsequently prepared an exhaustive mono- 
graph "On Apospory and Allied Phenomena". Linnean Transactions, 
vol. ii., part 14, July, 1887, to which reference should be made for details 
of the preceding cases. 


frond became covered with prothalli developed not merely 
from the edges, but also from the upper surface, and being- 
pegged down produced a number of plants, but whether 
apogamously or not I cannot say, though from De Bary's 
observations, they should be so. It is worthy of remark 
that in some of these youngsters, the line between the two 
generations of sporophore and oophore was so vague that 
the primary fronds were simply stalked prothalli, the next 
frond half one and half the other, while the fourth or 
fifth had quite outgrown the tendency and were of the 
typical varietal form. This plant was exhibited and de- 
scribed at the Linnean Society, 3rd November, 1892. 1 Of 
the next two cases I observed, the first was an Athyrium 
found in Lancashire and exhibited in 1893 at the meeting 
of the Pteridological Society at Lancaster by Mr. Bolton 
the finder. Immediately on seeing it I remarked, "How 
very like Col. Jones's Clarissima," simultaneously with which 
Mr. Bolton said, " It is strange, but it never ripens its 
spores " ". Turning the frond over, the reason was clear, 
it was perfectly white with aposporal excrescences. On 
submitting these to culture they produce plants freely by 
sexual action, but of two types, one very depauperate, mere 
skeleton plants, and the other of the parental form with 
occasional reversion towards the normal. In some of these 
young plants the whitish excrescences are plentiful in 
fronds only an inch or two high, and there are evident 
signs of prothalloid growth at the tips of the segments as 
well, pointing to apical apospory when the plants are more 
developed. The next case occurs in a most unlikely species, 
especially as apical apospory is in question. This is seen 
in a variety of Scolopendrium vulgare (S. v. cri spurn 
DrummondicB) which occurs in the wild state, like all the rest, 
characterised by being frilled and crested, and having more- 
over a finely fimbriated edge to the fronds. Visiting Mr. 
Bolton to inspect the Athyrium last cited, I saw a fine plant of 
this fern, and it immediately struck me that the tips of 

lu Notes on an Aposporous Lastrea (Nephrodium)" Linn. Soc. Journal 
Botany, vol. xxix., pp. 479-82. 


the fimbriate projections were remarkably translucent. I 
obtained material, laid it down, and at once prothalli began 
to develop vigorously from every point, so vigorously 
indeed that a single tip has formed a mass of prothalli an 
inch across which yielded at least a dozen plants of the 
parental form. 1 

Dr. F. W. Stansfield has recently sent me prothalli 
developed from a finely fimbriated form of Lastrea of 
which the reputed parent is that already described, and in- 
forms me that it is profusely aposporous though fairly de- 
veloped in size. 

By the various instances of this phenomenon so far 
cited, we see that the normal life cycles of the ferns in 
question have been successively shortened, first by the 
elision of the spore and then by that of the whole soral 
apparatus, while if we accept De Bary's observations as 
establishing the constant apogamous reproduction of L. 
pseudo mas cristata, in that case, it is shortened almost to 
the utmost, the chain being simply sporophore, prothallus, 
sporophore. Consistently indeed with the alternation of 
generation the chain could not apparently be shorter since 
the prothallus being eliminated we naturally come, or 
seem to come, to simple bulbils, such as occur on many 
ferns, Aspleniuvi bulbiferum for example. In the final 
case, however, which I have to cite, we arrive at the 
elimination even of the prothallus by substitution of the 
frond itself as the oophore or egg-bearer, the archegonia 
and antheridia being generated upon the frond without the 
prior formation of a prothallus proper. In a small plant 
of Scolopendrium vulgare recently sent me by Mr. E. J. 
Lowe, and exhibited by me at the Linnean Society in 
November last, although a definite axis of growth had been 
formed and several fronds had arisen in the normal spiral 
fashion around it, indicating that the prothallus stage had 
been unmistakeably passed, each of these fronds bore a 
thickened cushion at its tip upon which were seated both 

1<l Note on Apospory in a form of Scolopendrium vulgare" etc., Linn. 
Soc. Journal, vol. xxx., pp. 281-84. 


antheridia and archegonia, accompanied by aerial roothairs, 
the frond itself thus assuming the functions of the pro- 
thallus. Mr. Lowe raised a number of similar plants on 
the genesis of which he is preparing a paper which I will 
not forestall ; but he informs me that in time they throw 
off this aposporous character. Fronds which he has sent 
me, and which I have laid down, have developed prothalli 
all over their surface and at all terminals, but so far my 
cultures are too recent to permit me to report the advent of 

This completes the sketch of the cases which have 
come under my immediate notice, but considering that, in- 
cluding the first discovery, the phenomenon has been 
observed in no less than nine instances in our limited num- 
ber of British species, viz., Lastrea [Nephrodium) two, Athy- 
rium filix fcemina two, Polys tic hum angular e three, and 
Scolopendrium vulgare two ; it is only reasonable to ex- 
pect that many undiscovered instances must occur in the 
innumerable other species existent throughout the world. 

Charles T. Druery. 

Science progress* 

No. 28. 

June, 1896. 

Vol. V. 




IN the year 1868, spectrum analysis was first utilised in 
endeavouring" to unravel the message which was con- 
veyed to us by a most interesting eclipse observed in India. 
The diagrams will indicate the kind of record with which 
we have to deal in studying these celestial hieroglyphics. 
We are in one part dealing with the long waves of light, 
the red ; we are in the other dealing with the shorter waves 
of light, the blue. The work done in that eclipse is 
indicated by the bright lines — the hieroglyphics — which, 
when translated as they have been, describe for us the 
chemical nature of the particular stuff in the sun, which 
made him put on a blood-red appearance " on his getting 
out of his eclipse ". Taking the notes in the light scale 
which are lettered in the ordinary spectrum of sunlight, in 
order that they may be easily recognised and remembered, 
we learn the particular qualities of the light emitted by the 
blood-red streak. 

We have one quality represented by the line D, another 
at C, and another at F. According to the diagram, one of 
the lines is in the position of D. One observer said it was 
"at D, or near D ". 

Soon after this eclipse was observed in India, a method, 




long before formulated, of studying the blood-red streak 
surrounding the sun without waiting for an eclipse was 
brought into operation. 

By this method it was quite easy to make observations 
whenever the sun was shining, perfectly free from any of 
the difficulties attending the hurry and the worry and the 
excitement of an eclipse, which lasts only a few seconds. 

B C -D Ef F G 

Fig. i. — Pogson's diagram of the spectra of the sun's surroundings in the 
Eclipse of 1868. The bright lines seen are shown in the upper part 
of the diagram ; the chief lines in the solar spectrum, red to the left, 
blue to the right, are shown in the lower part. 

A 1 


D Eb 

F G 



. 1 


H/» Hr 




Fig. 2. — Summation of the observations of the spectrum of the 
sun's surroundings in the Eclipse of 1868. (1) Solar 
spectrum showing the position of the chief lines. (2) 
Rayet's observations of bright lines. (3) Herschel's obser- 
vations of bright lines. (4) Tennant's. 

Further, as the method consists of throwing an image of 
the sun, formed by a telescope, on to the slit of a spectro- 
scope, so that the spectrum of the sun's edge and of the 
sun's surroundings can be seen at the same time, exact 
coincidence or want of coincidence between the bright and 
dark lines can be at once determined. During an eclipse 


this of course is not possible, as the ordinary spectrum of 
the sun, with its tell-tale dark lines, is invisible because the 
sun, as we ordinarily see it, is hidden by the moon. 

Working, then, under such very favourable conditions, it 
was seen that there was certainly a red line given by this lower 



Fig. 3. — The exact coincidence of the red line with the dark line C. 

part of the solar atmosphere coincident with the very im- 
portant line in the solar spectrum which we call C. 

Another part of the spectrum in the blue-green was 
examined, and there again it was seen that the parts out- 
side the sun gave us a bright line exactly in the position of 

> ii 1 1 1 11 1 1 1 1 1 1 I ■ 

Fig. 4. — The exact coincidence of the blue-green line with the dark line F. 

the obvious dark line in the solar spectrum which is called 
F ; so that with regard to those two most important lines, 
there was no doubt whatever that we were dealing with 
the substance which produces these dark lines in the solar 



Fig. 5 is a diagram of the yellow, or rather the orange, 
part of the solar spectrum, showing two very important 
lines, which are called the lines D, due to the metal sodium, 
the investigation of which was just as important in solving 
the celestial hieroglyphics we call spectral lines as the 
Rosetta stone was important in settling the question of the 
Egyptian ones. 

Pogson, in referring to the eclipse of 1868, said that the 
orange line was "at D, or near D ". We see from the 

D 1 D 2 

Fig. 5. — The want of coincidence of the orange line D 3 with the dark 

lines D 1 and D 2 . 

diagram (Fig. 5) that the new method indicated that "near 
D " was the true definition. The line in this position in 
the spectrum, unlike the other two lines which I have 
indicated, has no connection at all with any of the dark 
lines in the ordinary solar spectrum. We were therefore 
perfectly justified in attaching considerable importance to 
this divergence in the behaviour of this line, taking the 
normal behaviour to be represented by the two strong lines 
in the red and the blue-green. The new line was called 
D 3 to distinguish it from the sodium lines D 1 and D 2 . 

A considerable amount of work was done with regard 
to the orange line. It was found that there was no sub- 
stance in our laboratories which could produce it for us, 
whereas in the case of the line D we simply had to burn 
some sodium, or even common salt, in a flame to produce 
it, and the other lines in the red and the blue-green were 
easily made manifest by just enclosing hydrogen in a 
vacuum tube, and passing an electric current through it, 


or observing the spectrum of a spark in a stream of coal- 

Now at the first blush it looked very much as if this 
line was really due to the same element which produced 
the others at C and F, and it was imagined that the reason 
we did not see it in our laboratories was because it was a 
line which required a very considerable thickness of hydro- 
gen to render it visible. That was the first idea, and Dr. 
Frankland and myself found that there was very consider- 
able justification for this view, because a simple calculation 
showed that the thickness of the solar atmosphere, which 
was producing that orange line under the conditions which 
enabled us to see it in our instruments by looking along the 
edge of the sun, was something like 200,000 miles. 

Fig. 6. — Changes of wave-length of the F hydrogen line when a solar 
cyclone is observed. A, the change towards the red indicates the 
retreating side of cyclone. C, the change towards the blue indicates 
the advancing side. B, the whole cyclone is included in the width of 
the slit, and both changes of wave-length are visible. 

Hence, in order to get a final decision on this point, 
there was nothing for it but to tackle the question from a 
perfectly different point of view, and the different point of 
view was this. The work had not gone on very long 
before one found minute alterations in the positions of these 
lines in the spectrum ; the orange line, for instance, might 
sometimes be slightly on one side, and sometimes on the 
other of its normal position. Further work showed that in 
these so-called " changes of wave-length " we had a precious 
means of determining the rate of movement of the gases 
and vapours in the solar atmosphere. 

Fig. 6 indicates how these changes of wave-lengths are 


shown in the spectroscope. The lines are contorted in both 
directions, and sometimes to a very considerable extent, 
indicating wind movements on the sun, reaching and some- 
times exceeding ioo miles a second. 

We had here a means of determining whether the 
orange line was produced by the same gases which gave the 
red and blue lines, because if so, when we got any altera- 
tion in the position of the red and blue lines, which always 
worked together, we should get an equivalent alteration in 
the position of the orange one. 

I found that the orange line behaved quite differently 
from either the red or the blue lines ; so then we knew that 
we were not dealing with hydrogen ; hence we had to do 
with an element which we could not get in our laboratories, 
and therefore I took upon myself the responsibility of coin- 
ing the word helium, in the first instance for laboratory 

This kind of work went on for a considerable time, and 
what one found was, that very often in solar disturbances 
we certainly were dealing with some of the lines of sub- 
stances with which we are familiar on this earth ; but at the 
same time it was very remarkable that when the records 
came to be examined, as they ultimately were with infinite 
care and skill, it was found that not only did we get this 
line in the orange indicating an unknown element associated 
with substances very well known, like magnesium, but that 
there were many other unknown lines as well. Within a 
few months of my first observations, several new lines about 
which nothing was known were thus observed. 


The place of the orange line D 3 I determined on 
20th October, 1868. Among many other lines behaving 
like it, two at wave-lengths 4923 and 5017 were discovered 
in June, 1869, and afterwards another at 6677, while Pro- 
fessor Young noted another in September, 1869, at 4471. 
He wrote : — 

" I desire to call special attention to 2581*5 [ = 4471 on 


Kirchhoff's scale], the only one of my list, by the way, 
which is not given on Mr. Lockyer's. This line, which was 
conspicuous at the Eclipse of 1869, seems to be always 
present in the spectrum of the chromosphere. . . . It has no 
corresponding dark line in the ordinary solar spectrum, and 
not improbably may be due to the same substance that 
produces D 3 ." 

This same line was noted also by Lorenzoni and named 
f. Another line at 4026 was added later by Professor 

Fig. 7. — Tacchini's observations of two slight solar disturbances 
showing the height to which the layers of the different gases 
extend. Magnesium vapour is highest of all, and is furthest 
extended ; next comes a gas of still unknown origin, indicated 
by a line at 1474 of Kirchhoff's scale and so on. 

Then with regard to solar disturbances. Let me refer 
in detail to a diagram indicating some results arrived at by 
the Italian observers. We are dealing with the spectro- 
scopic record of two slight disturbances in a particular part 
of the sun's atmosphere. The spectroscope tells us that in 
that region there was a quantity of the vapour of magnesium 
which is collected in that place. Then we find that another 
substance, about which we again know nothing whatever, 
is also visible in that region, and then we get the further 
fact that in those particular disturbances we get four other 
spectral lines indicated as being disturbed, and of those four 
lines we only know about one. 


In that way it very soon became perfectly clear to those 
who were working at the sun, that in all these disturbances, 
or at all events in most of them, we were dealing to a large 
extent with lines not seen in our laboratories when dealing 
with terrestrial substances ; this work went on till ultimately, 
thanks to the labours of Professor Young in America, we 
had a considerable list of lines coming from known and un- 
known substances which had been observed under these 
conditions in solar disturbances, and Professor Young was 
enabled to indicate the relative number of times these lines 
were visible. For instance, the lines which are most 
frequently seen under these conditions he tabulated as 
represented by the number 100, and of course the line 
which was least frequently seen would be represented by 
1 ; and therefore from these so-called "frequencies" we 
got a good idea of the number of times we might expect 
to see any of these disturbance-lines when anything was 
going on in the sun. 

It was this kind of work which made Tennyson write 
those very beautiful lines : 

" Science reaches forth her arms 
To feel from world to world ". 1 

1 And then he added : 

" and charms 
Her secret from the latest moon ". 

I mention this because Tennyson, whose mind was saturated with 
astronomy, had already grasped the fact that what had already been done 
was a small matter compared with what the spectroscope could do ; and 
now the prophecy is already fulfilled, for by means of the spectroscopic 
examination of the light from the stars we can tell that some of them are 
double stars, that is to say, in poetic language, stars with attendant moons. 
Although we can thus charm the secret from each moon by means of the 
spectroscope, to see the moon it would require (in the case of (3 Aurigse) a 
telescope not eighty feet long, but with an object-glass eighty feet in dia- 
meter, because the closer two stars are together the greater must be the 
diameter of the object-glass, independently cf its focal-length and magnifying 



In this year Dr. Hillebrand, one of the officials in the 
Geological Department at Washington, was engaged upon 
the chemical examination of specimens of the mineral 
uraninite from various localities. 

He dealt with crystals which he put in a vessel contain- 
ing some sulphuric acid and water. He found that bubbles 
of gas were produced out of the crystal by means of the 
sulphuric acid. He collected this gas and came to the 
conclusion that it was nitrogen. 

This result was new. He thus wrote about it : — 

"In consequence of a certain observation " [the one I 
have just referred to] " and its results, an entirely new 
direction was given to the work, and its scope wonderfully 
broadened. This was the discovery of a hitherto un- 
suspected element in uraninite, existing in a form of com- 
bination not before observed in the mineral world." 

It is not needful here to follow -Dr. Hillebrand through 
all the painstaking and patient labour he cut out for him- 
self to explain this anomalous behaviour. Needless to 
say he did not omit to employ the spectroscope to test the 
nature of the new gas. 

His observations were thus described : — 1 

"In a Geissler tube under a pressure of ten milli- 
metres and less, the gas afforded the fluted spectrum of 
pure nitrogen as brilliantly and as completely as was done 
by a purchased nitrogen tube. In order that no possibility 
of error might exist, the tube was then reopened and 
repeatedly filled with hydrogen, and evacuated till only the 
hydrogen lines were visible. When now filled with the 
gas and again evacuated, the nitrogen spectrum appeared 
as brilliantly as before, with the three bright hydrogen lines 

On this paragraph I may remark that it has long been 
known that gases like nitrogen give us quite distinct spectra 
at different temperatures — one fluted, another containing 

1<( On the Occurrence of Nitrogen in Uraninite," Bulletin, No. 78, 
U.S. Geol. Survey, 1889-90, p. 55. 


lines. Which of these we shall see in a tube will depend 
upon the pressure of the gas and the electric current used. 
The fluted spectrum of nitrogen is very bright and full of 
beautiful detail in the orange part of the spectrum ; the line 
spectrum, on the other hand, is almost bare in that region. 

It is important to note that it so happened 'that the pressure 
and electric conditions employed by Dr. Hillebrand enabled 
him generally to see the fluted spectrum. This however 
was not always the case. In an interesting letter to Pro- 
fessor Ramsay he writes (Proc. Roy. Soc, vol. lviii., p. 81): — 

" Both Dr. Hallock and I observed numerous bright 
lines on one or two occasions, some of which apparently 
could be accounted for by known elements — as mercury, or 
sulphur from sulphuric acid ; but there were others which I 
could not identify with any mapped lines. The well-known 
variability in the spectra of some substances under varying 
conditions of current and degree of evacuation of the tube 
led me to ascribe similar causes for these anomalous appear- 
ances, and to reject the suggestion made by one of us in a 
doubtfully serious spirit, that a new element might be in 

Dr. Hillebrand concludes his paper as follows : — 

"The interest in the matter is not confined merely to a 
solution of the composition of this one mineral ; it is broader 
than that, and the question arises : May not nitrogen be a 
constituent of other species in a form hitherto unsuspected 
and unrecognisable by our ordinary chemical manipulations? 
And, if so, other problems are suggested which it is not now 
in order to discuss." 


A negative of the nebula of Orion, taken at my observatory 
at Westgate-on-Sea in 1890, contains fifty-six lines, and of 
course by determining, as we have been able to do approxi- 
mately, the wave-lengths — the positions of these lines in 
the spectrum — we can determine the exact light notes 
represented, and therefore the substances which produce 
them. In this spectrum of the nebula of the Orion were 


lines of unknown origin exactly coinciding with those un- 
known lines which I have already referred to as having 
been seen in the sun's atmosphere. Some of the un- 
known lines in that atmosphere, those that we have not 
been able to see in our laboratories, are identical in position 
with some of the unknown lines in the nebula of Orion. 

I may remark that as early as 1886 Dr. Copeland had 
discovered D 3 in the visible spectrum of the nebula, and 
in a letter to him I had suggested that another line he had 
recorded at 447 might be Lorenzoni's f\ this he thought 
to be probable. The matter was set for ever at rest by 
the photograph which established the presence of 4471 and 
4026 as well, already noted as a solar line. 

Professor Campbell, of the Lick Observatory, obtained 
other photographs of the spectrum of the nebula some two 
or three years after mine was taken. In the following list 
of lines in my photograph an asterisk denotes that Campbell 
gives a line nearly in the same position. He recorded 
no line which did not appear on my photograph. 

401 1 







5875-8 = D3 

About the year 1890 I began the photography of stellar 
spectra at Kensington, with special reference to their 
classification on the basis of the chemical constituents 
established by their spectra. By 1892 several important 
results had been obtained, while the progress of this branch 
of science lately has been so considerable that any state- 
ment regarding the positions of lines, and therefore the 


chemical origins of them, may be made with a considerable 
amount of certainty as depending upon very accurate work. 

The various classes in which the stars have been 
classified by different observers according to their spectra 
are discussed elsewhere, but some of the more salient differ- 
ences must be pointed out here ; thus we have stars with 
many lines in their spectra, others with comparatively few. 
I will take the many-lined stars first. 

The diagram (Fig. 8) represents the spectrum of 
Arcturus, a star the spectrum of which closely resembles 
that of the sun. In a Cygni we have another star with 
many lines, but here we note, when we leave the hydrogen 
on one side and deal with the other stronger lines, that 
there is little relation between the solar spectrum and these 

I next come to the stars with few lines : these are well 
represented by many of the chief stars in the Constellation 
of Orion. Bellatrix is given as an example (Fig. 9). 

Then, I have next to say that in the photographs of the 
spectra of many stars, chiefly of those more or less like 
Bellatrix, we found the same lines which we have so far 
classified as unknown for the reason that in our laboratories 
we have not been able to get any lines which correspond 
with them. I again mention D 3 , 4471 and 4026, previously 
noted as appearing both in the chromosphere and in the 
nebula of Orion. 

But the thing is much more interesting even than this ; 
not only these, but all the chief unknown lines appearing in 
the nebula of Orion are also found in these stars. And this 
is so absolutely true that there is no necessity to give a list 
of the unknown lines seen in Bellatrix ; every one of them 
given in the nebula has found its place, and (so far) practically 
no others. 

This of course marked a great development of the 
inquiry, and makes the question of the unknown lines 
more important than ever. 












A method which was first employed by Respighi and 
myself during the eclipse of 1871, was employed on a 
large scale and with great effect during the eclipse of 1893. 
The light proceeding from the luminous ring round the 
dark moon was made to give us a series of rings, represent- 
ing each bright line seen by the ordinary method on a 
photographic plate. The observers this time were stationed 
in West Africa and in Brazil. The African station was 
up one of the rivers, not very far away from the town of 
Bathurst. The Brazilian station was near Para Curu. The 
same instrument which was previously referred to as used for 
obtaining photographs of the stars was sent to the African 
station in order that photographs of the eclipse of the sun 
might be taken on exactly the same scale as the photo- 
graphs of the stars had been, so that the stellar and solar 
records in the photographs might be compared. The results 
obtained by Messrs. Fowler and Shackleton, who were in 
charge of the instruments at the two stations, will be gathered 
from the accompanying diagrams, Figs. 10 and 11. 

We get more or less complete rings when we are deal- 
ing with an extended arc of the chromosphere, or lines of 
dots when any small part of it is being subjected to a dis- 
turbance which increases the temperature and, possibly, 
the numbers of the different vapours present. 

The efficiency of this method of work with the dis- 
persion employed turns out to be simply marvellous, and in 
securing such valuable and permanent records as these, we 
have done very much better than if we had contented our- 
selves with the style of observations that I have referred to 
as having been made in 1871. 

As was expected the comparison between solar and 
stellar records thus rendered possible enabled a very great 
advance to be made. 

On examining these eclipse records, we find that we 
have to do exactly with those unknown lines which had 
already been photographed in the stars and in the nebulas. 

As was to be expected we, of course, deal with the lines 



recorded in the first observations of the solar disturbances, 
and chronicled in that table of Professor Young's to which 
I have already called attention ; but the important thing is 
the marvellously close connection between eclipse- and star- 
spectrum photographs so far as the "unknown lines" are 

Nearly all the lines given in the table on p. 259 as 
visible in the Nebula of Orion and afterwards found in 
Bellatrix, are also among the lines photographed during the 


The year 1894 was made memorable by the announce- 
ment of the discovery by Lord Rayleigh and Professor 
Ramsay of a new gas called argon, and you know that the 
discovery was brought about chiefly in the first instance 
by the very accurate observations of Lord Rayleigh, who 
found that when he was determining the weight of air in 
the globe of a certain capacity, the weight depended upon 
the source from which he got the nitrogen. 

From the nitrogen from atmospheric air he obtained one 
weight, and from that obtained by certain chemical pro- 
cesses he obtained another, and ultimately it was found that 
there was an unknown element which produced these results, 
these various changes in the weight, and as a consequence 
we had the 1895 discovery of argon. 

Early in 1895 it struck Mr. Miers, of the British 
Museum, that it might be desirable to draw attention to 
the nitrogen which we have seen Dr. Hillebrand in 1888 
obtaining from his crystal of uraninite ; his observations, of 
course, were more in the mind of Mr. Miers than in the 
minds of the pure chemists. He therefore communicated 
with Professor Ramsay, who lost no time, because it was 
very interesting to study every possible source of nitrogen 
and see what its behaviour was in regard to the quantity 
of argon that it produced, and in the relation generally of the 
gas to the argon which was produced from it. 

Professor Ramsay treated uraninite in exactly the same 


C M 
O « 

S £ 
•55 o> 

t3 en 

O o 

u -a 
be c 

"g « 

i) U 


.c _ 


o »- 

<U to 
— <u 
— u 
55 C 
O u 


S S 

-= 2 
u a. 

4-* X 

c E 
.2 « 

o j= 


iL c 

U3 -4-* 

O 1) 


43 — 

■4-» *^ 

O ■"-> 
Ci 1— 1 



>- c/i 

4V D 


e E 

<u o 

4^ i: 

rt &, 



4-. <U 

2 E 

£ e.SP 
< .2^ 

M -T3 — 

a- c 

r! ■- 







2 - 

c/) 43 


J 9 


way that Dr. Hillebrand had done in 1888. The gas 
obtained as Dr. Hillebrand had obtained it was eventually- 
submitted to a spectroscopic test, following Dr. Hillebrand's 
example. But here a noteworthy thing comes in. 

It so happened that the pressure and electrical conditions 
employed by Professor Ramsay were so different from those 
used by Dr. Hillebrand that, although nitrogen was un- 
doubtedly present, the fluted spectrum which, as I have 
previously stated, floods the orange part of the spectrum 
with luminous details, was absent. But still there was 
something there. 

Judge of Professor Ramsay's surprise when he found 
that he got a bright orange line ; that was the chief thing, 
and not the strong suggestion of the spectrum of nitrogen. 
Careful measurements indicated that the twenty-six-year- 
old helium had at last been run to earth, D 3 was at last 
visible in a laboratory. Professor Ramsay was good enough 
to send specimens of the tubes containing this gas round 
to other people, and he sent one of them to me. 

I received Professor Ramsay's tube on 28th March, but 
it was not suitable for the experiments I wished to make. 

On 29th March, therefore, as Professor Ramsay was 
absent from England, in order not to lose time I determined 
to see whether the gas which had been obtained by chemical 
processes would not come over by heating in vacuo, after the 
manner described by me to the Royal Society in 1879/and Mr. 
L. Fletcher was kind enough to give me some particles of 
uraninite (broggerite) to enable me to make the experiment. 

This I did on 30th March, and it succeeded ; the gas 
giving the yellow line came over, associated with hydrogen, 
in good quantity. 

From 30th March onwards my assistants and myself 
had a very exciting time. One by one the unknown lines 
I had observed in the sun in 1868 were found to belong to 
the gas I was distilling from broggerite ; not only D 3 but 
4923, 5017, 4471 (Lorenzoni's/), 6677 (the B C of Fig. 7), 
referred to previously, and many other solar lines, were all 
caught in a few weeks. 

1 Roy. Soc. Proc, vol. xxix., p. 266. 


But this was by no means all. The solar observations 
had been made by eye, and referred therefore to the less 
refrangible part of the spectrum, but I had obtained and 
studied hundreds of stellar photographs, so I at once pro- 
ceeded to photograph the gas and compare its more re- 
frangible lines with stellar lines. 

Here, if possible, the result was still more marvellous. 
In the few-lined stars by 6th May I had caught nearly all 
the most important lines at the first casts of the spectroscopic 
net. Fig. 1 5, which includes some later results, will give an 
idea of the tremendous revelation which had been made as 
to the chemistry of some of the stages of star-life. I 
pointed out on 8th May that we had already "run home " 
the most important lines in the spectra of Group III. in 
which stars alone we find D 3 reversed. 

These results enabled us at once to understand how it 
was that the "unknown lines" had been seen both in the 
sun's chromosphere and some nebulae and stars. The gas 
obtained from the minerals made its appearance in the 
various heavenly bodies in which the conditions of the 
highest temperatures were present ; and the more the work 
goes on, we find that this gas is really the origin of most, 
but certainly not of all, of the unknown lines which have 
been teasing astronomical workers for the last quarter of a 



The dates of the papers communicated to the Royal 
Society recording the observations of the lines in the gas 
obtained from minerals which had been previously recorded 
are as follows : — 



The lines at 667 and 5016 had been previously seen 
by Thalen (Comptes Rendus, 16th April, 1895). 

25 th April, - 

- 447i 


8th May, 

- 667 


9th May, 

- 3889 

28th May, - 

- 7065 

29th May, - 

- 5 48 



Although the general distribution and intensities of the 
lines in the gases from broggerite and cleveite sufficiently 
corresponded with some of the chief " unknown lines " in 
the solar chromosphere and some of the stars to render 
identity probable, it was necessary to see how far the con- 
clusion was sustained by detailed investigations of the 
wave-lengths of the various lines. 


This was practically a separate branch of the work, as 
the observations had to be made in the observatory. Next 
I give here the observations relating- to D 3 , 4471. 

The Orange Line, A 5875*9. — Immediately on receiving 
from Professor Ramsay, on 28th March, a small bulb of the 
gas obtained from cleveite, a provisional determination of 
wave-length was made by Mr. Fowler and myself, in the 
absence of the sun, by micrometric comparisons with the D 
lines of sodium, the resulting wave-length being 5876*07 
on Rowland's scale. It was at once apparent, therefore, 
that the gas line was not far removed from the chromo- 
spheric D 3 , the wave-length of which is given by Rowland 

as 5875'98. 

The bulb being too much blackened by sparking to give 
sufficient luminosity for further measurements, I set about 
preparing some of the gas for myself by heating broggerite 
in vacuo, in the manner I have already described. A new 
measurement was thus secured on 30th March, with a 
spectroscope having a dense Jena glass prism of 6o° ; this 
gave the wave-length 5876*0. 

On 5th April, I attempted to make a direct comparison 
with the chromospheric line, but though the lines were 
shown to be excessively near to each other, the observa- 
tions were not regarded as final. 

Professor Ramsay having been kind enough to furnish 
me, on 1st May, with a vacuum tube which showed the 
orange line very brilliantly, a further comparison with the 
chromosphere was made on 4th May. The observations 
were made by Mr. Fowler, in the third order spectrum of 
a grating having 14,438 lines to the inch, and the observing 


telescope was fitted with a high power micrometer eye- 
piece ; the dispersion was sufficient to easily show the 
difference of position of the D 3 line on the east and west 
limbs, due to the sun's rotation. Observations of the 
chromosphere were therefore confined to the poles. 

During the short time that the tube retained its great 
brilliancy, a faint line, a little less refrangible than the 
bright orange one, and making a close double with it, was 
readily seen ; but afterwards a sudden change took place, 
and the lines almost faded away. While the gas line was 
brilliant, it was found to be " the least trace more refrangible 
than D 3 , about the thickness of the line itself, which was 
but narrow" ("Observatory Note Book"). The sudden 
diminution in the brightness of the lines made subsequent 
observations less certain, but the instrumental conditions 
being slightly varied, it was thought that the gas line was 
probably less refrangible than the D 3 line by about the 
same amount that the first observation showed it to be 
more refrangible. Giving the observations equal weight, 
the gas line would thus appear to be probably coincident 
with the middle of the chromospheric line, but if extra 
weight be given to the first observation, made under much 
more favourable conditions, the gas line would be slightly 
more refrangible than the middle of the chromosphere line. 

Pressure of other work did not permit the continuation 
of the comparisons. In the meantime, Runge and Paschen 
announced (Nature, vol. Hi., p. 128) that they also had seen 
the orange line of the cleveite gas to be a close double, 
neither component having exactly the same wave-length as 
D 3 , according to Rowland. 

They give the wave-length of the brightest component as 
5878*883, and the distance apart of the lines as 0*323. 

This independent confirmation of the duplicity of the 
gas line led me to carefully re-observe the D 3 line in the 
chromosphere for evidences of doubling. On 14th June 
observations were made by Mr. Shackleton and myself of 
the D 3 line in the third and fourth order spectra under 
favourable conditions ; " the line was seen best in the fourth 
order, on an extension of the chromosphere or prominence 


on the north-east limb of the sun. The D 3 line was seen 
very well, having every appearance of being double, with a 
faint component on the red side, dimming away gradually ; 
the line of demarcation between the components was not 
well marked, but it was seen better in the prominence than 
anywhere else on the limb " (" Observatory Note Book "). 

It became clear, then, that the middle of the chromo- 
sphere line, as ordinarily seen, and as taken in the 
comparison of 4th May, does not represent the place of 
the brightest component of the double line, so that exact 
coincidence was not to be expected. 

The circumstance that the line is double in both gas 
and chromosphere spectrum, in each the less refrangible 
component being the fainter, taken in conjunction with the 
direct comparisons which have been made, rendered it 
highly probable that one of the gases obtained from cleveite 
is identical with that which produces the D 3 line in the 
spectrum of the chromosphere. 

Other observers have since succeeded in resolving the 
chromospheric line. On 20th June, Professor Hale found 
the line to be clearly double in the spectrum of a promin- 
ence, the less refrangible component being the fainter, and 
the distance apart of the lines being measured as 0*357 
tenth -metres (Ast. JVac/i., 3302). 

The doubling was noted with much less distinctness in 
the spectrum of the chromosphere itself on 24th June. 
Professor Hale points out that Rowland's value of the wave- 
length (as well as that of 5875*924, determined by himself 
on 19th and 20th June) does not take account of the fact 
that the line is a close double. 

Dr. Huggins, after some failures, observed the D 3 line 
to be double on 10th July [Ast. Nack., 3302); he also 
notes that the less refrangible component was the fainter, 
and that the distance apart of the lines was about the same 
as that of the lines in the gas from cleveite, according to 
Runge and Paschen. 

It may be added, that in addition to appearing in the 
chromosphere, the D 3 line has been observed as a bright 
line in nebulae by Dr. Copeland, Professor Keeler and 


others ; in /3 Lyrae and other bright line stars ; and as a 
dark line in such stars as Bellatrix, by Mr. Fowler, Pro- 
fessor Campbell and Professor Keeler. In all these cases 
it is associated with other lines, which, as I shall show pre- 
sently, are associated with it in the spectra of the new gases. 

The Blue Line, A 4471*8. — A provisional determination 
on 2nd April of the wave-length of a bright blue line, seen 
in the spectrum of the gases obtained from a specimen of 
cleveite, showed that it approximated very closely to a 
chromospheric line, the frequency of which is stated as 100 
by Young. 

This line was also seen very brilliantly in the tube 
supplied to me by Professor Ramsay on 1st May, and on 
6th May it was compared directly with the chromosphere 
line by Mr. Fowler. The second order grating spectrum 
was employed. The observations in this region were not 
so easy as in the case of D 3 , but with the dispersion em- 
ployed, the gas line was found to be coincident with the 
chromospheric one. In this case also, the chromosphere 
was observed at the sun's poles, in order to eliminate the 
effects due to the sun's rotation. 

Besides appearing in the spectrum of the chromosphere, 
the line in question is one of the first importance in the 
spectra of nebulae, bright line stars, and of the white stars 
such as Bellatrix and Rigel. 

The Infra-red Line, \ 7065*5. — In addition to D 3 and 
the line at 447 1 '8, there is a chromospheric line in the infra- 
red which also has a frequency of 100, according to Young. 
On 28th May I communicated a note to the Royal Society 
stating that this line had been observed in the spectrum of 
the gases obtained from broggerite and euxenite {Roy. Soc. 
Proc, vol. lviii., p. 192), solar comparisons having con- 
vinced me that the wave-length of the gas line corresponded 
with that given by Young ; and I added : " It follows, there- 
fore, that besides the hydrogen lines all three chromospheric 
lines in Young's list which have a frequency of 100 have 
now been recorded in the spectra of the new gas or gases 
obtained from minerals by the distillation method ". 

M. Deslandres, of the Paris Observatory, has also 


observed the line at 7065 in the gas obtained from the 
cleveite (Comptes Rendus, 17th June, 1895, p. 1 331). 

A great deal of work has been done upon these gases 
from other points of view than those which affect their 
cosmical relations, and perhaps I may be allowed next to 
refer to some of the results which have been obtained by 


The first point is that the gas from the minerals contains 
no argon. Dr. Ramsay in his first experiments came to the 
conclusion that the spectra of argon and helium contained 
many common lines ; indeed at first the observed coin- 
cidences were so remarkable that he came to the conclusion 
that the connection was so close that atmospheric argon con- 
tained a gas absent from the argon seen in his helium tube. 

This statement was subsequently withdrawn, but the 
compound nature both of argon and helium was suggested 
by the fact that there were lines common to the two gases. 
These lines were in the red ; one coincidence I found broke 
down with moderate dispersion, the other yielded subse- 
quently to the still greater dispersion employed by Drs. 
Runge and Paschen. It may be also stated here that I have 
not found a single coincidence between argon and any line 
in the spectrum of any celestial body whatever. This 
happens, as everybody knows, also in the case of oxygen, 
nitrogen, chlorine, and the like. 


The first spectroscopic observations made it perfectly 
obvious that the gas as obtained from uraninite is a mixture 
of gases, that the gas which gives the yellow line is not an 
isolated one, but is mixed up with other gases which give 
other lines. 

In May I wrote as follows : — 1 

" The preliminary reconnaissance suggests that the gas 
obtained from broggerite by my method is one of complex 

1 Proc. R. S., Iviii., p. 114. 


" I now proceed to show that the same conclusion holds 
good for the gases obtained by Professors Ramsay and 
Cleve from cleveite. 

" For this purpose, as the final measures of the lines of 
the gas as obtained from cleveite by Professors Ramsay and 
Cleve have not yet been published, I take those given by 
Crookes and Cleve, as observed by Thalen. 

" The most definite and striking result so far obtained is 
that in the spectra of the minerals giving the yellow line 
I have so far examined, I have never once seen the lines 
recorded by Crookes and Thalen in the blue. This demon- 
strates that the gas obtained from certain specimens of 
cleveite by chemical methods is vastly different from that 
obtained by my method from certain specimens of brog- 
gerite, and since, from the point of view of the blue lines, 
the spectrum of the gas obtained from cleveite is more 
complex than that of broggerite, the gas itself cannot be 
more simple. 

" Even the blue lines themselves, instead of appearing 
en bloc, vary enormously in the sun, the appearances being 

4922 (4921-3) = thirty times 
4713 (47 1 2-5) = twice. 

" These are not the only facts which can be adduced to 
suggest that the gas from cleveite is as complex as that 
from broggerite." 

It is seen that quite early in the inquiry we had not only 
spectroscopic evidence in the laboratory which was com- 
plete in itself, but that the case was greatly strengthened 
when the behaviour of the various lines in the sun and stars 
was also brought into evidence. 

In the first case we had the laboratory separation of D, 
from the lines 5048, 5016, and 4922. 

Later on in the same month I showed that the lines at 
D 3 and 447 behaved in one way, and that at 667 behaved 
in another. 

In order to test this view I made some observations 
based on the following considerations : — 

(1) In a simple gas like hydrogen, when the tension of 
the electric current given by an induction coil is increased 


by inserting first a jar and then an air-break into the circuit, 
the effect is to increase the brilliancy and the breadth of all 
the lines, the brilliancy and breadth being greatest when the 
longest air-break is used. 

(2) Contrariwise, when we are dealing with a known 
compound gas ; at the lowest tension we may get the 
complete spectrum of the compound without any trace of 
its constituents, and we may then, by increasing the tension, 
gradually bring in the lines of the constituents, until, when 
complete dissociation is finally reached, the spectrum of the 
compound itself disappears. 

Working on these lines the spectrum of the spark at 
atmospheric pressure passing through the gas or gases, 
distilled from broggerite, has been studied with reference 
to the special lines C (hydrogen), D 3 , 667, and 447. 

The first result is that all the lines do not vary equally 
as they should do if we were dealing with a simple gas. 

The second result is that at the lowest tension 667 is 
relatively more brilliant than the other lines ; on increasing 
the tension C and D 3 considerably increase their brilliancy, 
667 relatively and absolutely becoming more feeble, while 
447, seen easily as a narrow line at low tension, is almost 
broadened out into invisibility as the tension is increased 
in some of the tubes, or is greatly brightened as well as 
broadened in others (Fig. 12). 


D 3 



6563 667. 



Fig. 12. — Diagram showing changes in intensities of lines brought about by varying the 
tension of the spark, i. Without air-break. 2. With air-break. 

The above observations were made with a battery of 
five Grove cells ; the reduction of cells from 5 to 2 made 
no difference in the phenomena except in reducing their 

Reasoning from the above observations it seems evident 
that the effect of the higher tension is to break up a com- 
pound or compounds, of which C, D 3 , and 447 represent 
constituent elements ; while, at the same time, it would 


appear that 667 represents a line of some compound which 
is simultaneously dissociated. 

The unequal behaviour of the lines has been further 
noted in another experiment, in which the products of 
distillation of broggerite were observed in a vacuum tube 
and photographed at various stages. After the first heating 
D 3 and 447 1 were seen bright, before any lines other than 
those of carbon and hydrogen made their appearance. 
With continued heating 667, 5016, and 492 also appeared, 
although there was no notable increase of brightness in the 
yellow line ; still further heating introduced additional lines, 
5048 and 6347. 

These changes are represented graphically in the fol- 
lowing diagram (Fig. 13). 

D 3 

447. 492.501. 5876. 634 667. 

Fig. 13. — Diagram showing order in which lines appear in spectrum of vacuum tube 

when broggerite is heated. 

It was recorded further that the yellow line was at times 
dimmed, while the other lines were brightened. 

In my second note, communicated to the Royal Society 
on the 8th May, I stated that I had never once seen the 
lines recorded by Thalen in the blue, at A 4922 and 4715. 

It now seems possible that their absence from my 
previous tubes was due to the fact that the heating of the 
minerals was not sufficiently prolonged to bring out the 
gases producing these lines. 

It is perhaps to the similar high complexity of the gas 
obtained from cleveite that the curious behaviour of a tube 
which Professor Ramsay was so good as to send me, must 
be ascribed. When I received it from him the glorious 
yellow effulgence of the capillary while the current was 
passing was a sight to see. But after this had gone on 
for some time, while the coincidence of the yellow line with 
D 3 of the chromosphere was being inquired into, the lumi- 
nosity of the tube was considerably reduced, and the colours 


in the capillary and near the poles were changed. From 
the capillary there was but a feeble glimmer, not of an 
orange tint, while the orange tint was now observed near 
the poles, the poles themselves being obscured by a coating 
on the glass of brilliant metallic lustre. 

After attempting in vain for some time to determine the 
cause of the inversion of D 3 and 447 in various photographs 
I had obtained of the spectra of the products of distillation 
of many minerals, it struck me that these results might be 
associated with the phenomena exhibited by the tube, and 
that one explanation would be rendered more probable if it 
could be shown that the change in the illumination of the tube 
was due to the formation of platinum compounds, platinum 
poles being used. On 2 1st May I accordingly passed the cur- 
rent and heated one of the poles, rapidly changing its direction 
to assure the action of the negative pole, when the capillary 
shortly gave a very strong spectrum of hydrogen, both lines 
and structure. A gentle heat was continued for some time, 
and apparently the pressure in the tube varied very con- 
siderably, for as it cooled the hydrogen disappeared and the 
D 3 line shone out w T ith its pristine brilliancy. The experi- 
ment was repeated on 24th May, and similar phenomena 
were observed. 

Some little time after 1 Professors Runge and Paschen, 
from an entirely different standpoint, arrived at exactly the 
same conclusion. 

The employment of exposures extending over seven 
hours has given a considerable extension in the number of 
lines, and the bolometer has been called in to investigate 
lines in the infra-red ; better still, they have employed well- 
practised hands in searching for series of lines. Operating 
by chemical means upon a crystal of cleveite free from any 
other mineral, they have obtained a product so pure that 
from these series there are no outstanding lines. Very 
great weight, therefore, must be attached to their conclusions. 

As a result of their investigations Drs. Runge and 
Paschen stated that the gas given off even by a pure crystal 

1 Nature, 26th September, 1895. 


of cleveite is not simple. In their view the mixture consists 
of two constituents. 

This conclusion was arrived at from the following con- 
siderations. " The wave-lengths A of the lines belonging to 
the same series are always approximately connected by a 
formula somewhat similar to Balmer's — 

i/X = A - B/;// 2 - C/m\ 

A determines the end of the series towards which the lines 
approach for high values of m, but does not influence the 
difference of wave-numbers of any two lines. B has nearly 
the same value for all the series observed, and C may be 
said to determine the spread of the series, corresponding 
intervals between the wave-numbers being larger for larger 
values of C. As B is approximately known two wave- 
lengths of a series suffice to determine the constants A and 
C, and thus to calculate approximately the wave-lengths of 
the other lines. It was by this means that we succeeded in 
disentangling the spectrum of the gas in cleveite, and 
showing" its regularity. 

"In the spectrum of many elements two series have been 
observed for which A has the same value, so that they both 
approach to the same limit. In all these cases the series 
for which C has the smaller value, that is to say, which has 
the smaller spread, is the stronger of the two. In the 
spectrum of the gas in cleveite we have two instances of 
the same occurrence. One of the two pairs of series, the 
one to which the strong yellow double line belongs, consists 
throughout of double lines whose wave-numbers seem to have 
the same difference, while the lines of the other pair of series 
appear to be all single. Lithium is an instance of a pair of 
series of single lines approaching to the same limit. But 
there are also many instances of two series of double lines 
of equal difference of wave-numbers ending at the same 
place as sodium, potassium, aluminium, etc. There are also 
cases where the members of each series consist of triplets of 
the same difference of wave-numbers, as in the spectrum of 
magnesium, calcium, strontium, zinc, cadmium, mercury. 
But there is no instance of an element whose spectrum 
contains two pairs of series ending at the same place. This 


suggested to us the idea that the two pairs of series belonged 
to different elements. One of the two pairs being by far 
the stronger, we assume that the stronger one of the two 
remaining series belongs to the same element as the stronger 
pair. We thus get two spectra consisting of three series 
each, two series ending at the same place, and the third 
leaping over the first two in large bounds and ending in the 
more refrangible part of the spectrum. This third series we 
suppose to be analogous to the so-called principal series in 
the spectra of the alkalis, which show the same features. 
It is not impossible, one may even say not unlikely, that 
there are principal series in the spectra of the other elements. 
But so far they have not been shown to exist. 

" Each of our two spectra now shows a close analogy to 
the spectra of the alkalis. 

"We therefore believe the gas in cleveite to consist of 
two, and not more than two, constituents." 

To the one containing the line D 3 , which I discovered 
in 1868, the name helium remains ; the other for the present 
we may call " gas X V 

The chief lines of these two constituents are as follows, 
according to Runge and Paschen, the wave-lengths being 
abridged to five figures. 


Hi I ol 

I II] L 

If ' 

i ! 


r 5 



, r in. 


j A 

• .1 






J _ 








"** ' 


Fig. 14. — Runge and Paschen's results suggesting that cleveite gives off 
two gases, each with three series of lines. 

1 In the many comparisons I had to make, I soon found the incon- 
venience of not having a name for the gas which gave 667, 501 and other 
lines. When, therefore, Professors Runge and Paschen, who had endorsed 



1st Subordinate 

2nd Subordinate 

Principal Series. 
















3479' 1 









35 I2 '6 
















1st Subordinate 

and Subordinate 

Principal Series. 












3 2 3 I *3 























my results, and had extended them, called upon me, I thought it right to 
suggest to them that, sinking the priority of my own results, we should all 
three combine in suggesting a name. Professor Runge (under date 20th 
October) wrote me : " The inference that there are two gases is a spectro- 
scopical one, being based on the investigation of the ' series '. Now, though 
we think this basis quite sound, we must own that the conclusion rests on 
induction. . . . For this reason we do not want to give a name to ' gas 
X '." I have so far suggested no name, though Orionium and Asterium 
have been in my mind. 



More recently Professor Ramsay has abandoned his 
view of the simple nature of the cleveite gas, and states 
that from his experiments "there appears ground for the 
supposition that helium is a mixture ". 1 


And now comes the great revelation, and it is this. 
The majority of the lines classed as unknown in the spectra 
of the Orion nebula, stars of Group III. and the sun are 
really due to the cleveite gases. 

The following table sets this result out. It will be seen 
that of seventeen unknown lines, twelve have been run to 


Orion Nebula. 

Bellatrix and 
Eclipse, 1893. 





* 3 86 9 (7) 





3888 (7) 




4°i 1 (3) 

4009 (8; 



4026 (5) 

4026 (10) 



4042 (1) 

4041 (3) 

Still Unknown 


4068 (3) 

4°7° (3) 

Still Unknown 


4121 (1) 




4143 ( T ) 

4144 (8) 



4168 (1) 

4169 (5) 



4270 (3) 

4268 (7) 

Still Unknown 


439° (3) 

4389 (8) 



447 2 (7) 

4472 (10) 



454° (3) 

454i (1) 

Still Unknown 


4628 (3) 

4630 (3) 

Still Unknown 


47i6 (3) 

47i5 (5) 


■ — 

*4924 (5) 

f4922 - i (8) 





D 3 He. 

* Between these AA there are forty-two lines in the Orion photograph of which six are 
known other than He. and X. 

t Between these AA there are forty-five lines in the Bellatrix photograph of which 
five are known other than He. and X. 

The following tables give the complete list of lines 
and the celestial body in which they have been traced. 

1 Nature, vol. liii., p. 598. 


In the tables, under "sun," C, followed by a number, 
indicates the frequency as given by Young ; E indicates the 
lines photographed during the eclipse of 1893. Under 
"star or nebula" the references are to the tables given in 
my memoir on the nebula of Orion {Phil. Trans., vol. 
clxxxvi., 1895, P- 86 et seq. N = Nebula of Orion). 


I 1220. 


Star or Nebula. 


C E 

N. III. y 










C 100 E 


C 100 E 


C 25 E 



a Cygni 


3 6 34 



35 I 3 







346i y 


C 100 


C 2 E 



N. a Cygni 















* Means that these lines are out of the range of my observations. 






Star or Nebula. 


C 30 E 



III. y 

3 6 M> 




3 2 97 



323 1 

32i3 J 


C 25 


C 30 E 



N. III. y 



III. y 


III. y 







Hid byH. line 






C 2 







N. III. y 

393 6 

Hid in K. 


C E 

a Cygni 


C E 

a Cygni 


* Means that these lines are out of the range of my observations. 

The annexed reproduction of a photograph of Bellatrix 
will show how striking has been the result of the discovery 
so far as stellar spectra are concerned. 

Hydrogen, helium and gas X are thus proved to be 
those elements which are, we may say, completely repre- 
sented in the hottest stars and in the hottest part of the 
sun that we can get at. Here then, in 1895, we have 
abundant confirmation of the views I put forward in 
1868 as to the close connection between helium and 



A diffusion experiment described in their paper enabled 
Messrs. Runge and Paschen to go a stage farther, and to 













- -!: " 









; ... _ 




•;'■•■" HhSFv 





; > 







.• • ^S^f&v^i 











cr 3* 



r* rt 

p ,_. 


D- tzi 



1 ■ 


- iy: '-y;~; y'y-^. 


3 g- 


; .,.''■ 



SL 3 



Cfl ^ 

" 3 







,- ■ 








> z* 


• - 

— U. ■ - 

447 2 



I • 







'". -! 

r:^,#*i^f' : -'i 













» - ! •'. 










£$£a, •' "- 








<-$ w w TJ 

announce that of their two constituents the gas-giving D 3 
was the heavier one. They also add :— 

" From the fact that the second set of series is on the 


whole situated more to the refrangible part of the spectrum, 
one may, independently of the diffusion experiment, con- 
clude that the element corresponding to the second set is the 
heavier of the two ". 

As they themselves pointed out, however, the result was 
not final, because the pressures were not the same. I have 
recently made some experiments in which the pressures 
remain the same. 

An U tube was taken, and at the bend was fixed a 
plaster of Paris plug about 1*5 cm. thick; in one of the 
limbs two platinum wires were inserted. The plug was 
saturated with hydrogen to free it from air ; the tube was 
then plunged into a mercury trough, and fixed upright with 
the limbs full of mercury. Into the leg (A) with the plati- 
num wires a small quantity of hydrogen was passed, and as 
soon after as possible another small quantity of a mixture 
of helium and hydrogen from samarskite was put up the 
other limb (B) of the U tube. 

Immediately after the helium was passed into the limb 
(B) spectroscopic observations were made of the gas in the 
limb (A) ; D 3 was already visible, and there was no trace of 
50157. This result seems to clearly indicate that if a true 
diffusion of one constituent takes place, the component which 
gives D 3 is lighter than the one which gives the lines at 
wave-length 50157. 

Although this result is opposed to the statement made 
by Runge and Paschen, it is entirely in harmony with the 
solar and stellar results. 

In support of this I may instance that of the cleveite 
lines associated with hydrogen in the chromosphere and the 
stars of Group III. y ; those allied to D 3 are much stronger 
than those belonging to the series of which 50157 forms 


So far I have worked upon some seventy minerals, and 
I have found the orange line in sixteen. 

The following are the minerals, etc., which have been 


investigated ; those which give the D. line beino- marked 
with an asterisk : — 


































Manganese Nodule. 







Plumbic Ochre. 


Red Clay. 











J. Norman Lockyer. 


PART VI. (a). 

IN the preceding articles I have briefly reviewed the 
literature relating to Insular Floras which has appeared 
during the last decade, and I have extracted therefrom the 
principal or most interesting facts, which I have given with 
some comments of my own. That I have been able to do 
this with some profit is largely due to the advantages I have 
enjoyed through the kindness of the Director of the Royal 
Gardens, Kew. Since the publication in 1885 of my first 
essay on this subject, in the Botany of the Voyage of H. M.S. 
" Challenger" all or nearly all collections of insular plants 
received at Kew have passed through my hands for determina- 
tion and reporting on ; and I have also been favoured with 
many notes and criticisms by travellers and other persons 
interested in plant distribution. I propose therefore to 
enter into a short recapitulation and discussion of the main 
facts thus accumulated ; but before doing - so I will refer to 
some more or less important contributions to the subject 
that have come to light during the progress of the present 
series of articles. 1 

It will be convenient to take the islands in the same 
geographical order previously followed (1), beginning with 

There are some interesting recent contributions to the 
flora of Polynesia, taking the designation in its widest 
sense ; but no one has yet attempted to bring together what is 
known, or ascertainable from materials preserved in herbaria, 
of the vegetation of the numerous small coral islands and 
groups of islands, more or less recently annexed by, or taken 
under the protection of, Great Britain. This the writer is 
engaged upon, and some particulars thus acquired may be 

1 A review of the additional literature having extended beyond what 
was expected, the discussion referred to will form the subject of a conclud- 
ing article. 


utilised here in dealing with the literature of the subject. 
Some years ago Mr. J. T. Arundel delivered a lecture at 
San Francisco, before the Geographical Society of the 
Pacific, on the Phcenix Group and other islands of the 
Pacific, and he has since published it (2) with additional 
notes. Mr. Arundel writes from actual experience, having 
visited a large number of the most remote islets of the 
Pacific and collected samples of their scanty floras, which 
were determined for him at Kew, where the specimens are 
preserved. Unfortunately several of the names of the 
plants in his list have undergone such a transformation as 
to be almost unrecognisable. 

Besides the Phcenix Group, which was under his personal 
control, Mr. Arundel visited such out-of-the-way islands as 
Starbuck, Caroline (not the Caroline Group), Fanning, 
Maiden, Palmerston and Ducie. Mr. Arundel describes 
Starbuck and Caroline Islands as examples of two kinds of 
very small islands common in the Pacific, though not con- 
fined to it. The former represents those consisting of an 
unbroken mass which is treeless, and indeed almost devoid 
of vegetation ; and the latter is a typical coral atoll, con- 
sisting of a ring of islets encircling a central lagoon, and 
supporting a relatively luxuriant vegetation. Starbuck is 
very scantily furnished with vegetation, only about half a 
dozen species being represented. The principal plants are 
Lepidium piscidium and Sida fallax ; both of wide range 
in Polynesia. Caroline Island claims a little more atten- 
tion, because its history, position, conformation, meteorology, 
botany and zoology have been very fully worked out and 
illustrated. In 1883 this island was selected by the Ameri- 
cans, by the British, and by the French as the most suitable 
spot for observing the total eclipse of the sun. The Ameri- 
can party was relatively numerous, and they drew up a 
somewhat elaborate report (3), illustrated chiefly by prints 
from photographs taken by the two gentlemen constituting 
the English party. These illustrations give an excellent 
idea of the form and vegetation of an atoll, including a 
bird's eye view, which enables us, better than any description 
could, to realise its smallness and isolation. Caroline Island 


is situated in almost exactly 150° W. longitude and io° S. 
latitude, and is distant, according to Mr. Arundel, about 
400 miles from Tahiti, the nearest island of considerable 
size — say a third larger than the Isle of Wight ; and 420 
from Starbuck. Although in most parts well clothed with 
vegetation, this vegetation consists of very few, perhaps 
not more than twenty, species of vascular plants. Several 
others now exist, either as the remains of cultivation or 
accidental introduction ; and the abundance of the cocoanut 
palm is due to planting, which has now been in operation 
for some years. Whether the cocoanut existed in the 
island on the first advent of man there is no evidence to 
show ; but there are trees of other kinds of large size, as 
depicted and described in the report referred to. They are: 
Calophyllum Inophylliim (Guttiferse), Morinda citrifolia 
(Rubiaceae), Cordia subcordata (Boragineae), Pisonia grandis 
(Nyetaginaceae), and a screw pine, probably the widely spread 
Pandanus odoratissi?nus. One of the illustrations is a most 
effective representation of a group of screw pines. The 
Cordia is perhaps the commonest tree, and is most con- 
spicuous, having a spreading crown with branches down to 
the ground. Pisojiia grandis is described as forty or fifty 
feet high, with a trunk four feet in diameter ; dimensions 
one would hardly have expected. I have drawn some- 
what freely from this report, because it is by far the most 
instructive known to me. 

A more recent contribution to island literature by Mr. 
C. M. Woodford (4) is equally deserving of attention, 
though wanting illustrations. It deals with the Gilbert 
Archipelago, one of the most remarkable of the numerous 
groups in the Eastern Pacific. There are sixteen islands, 
not counting the islets of the atolls, forming a chain, trend- 
ing from north-west to south-east and extending from about 
3° north to 3 south latitude in 173 to 1 77° east longitude. 
Eleven out of the sixteen are of atoll formation, and the 
largest of them is little more than twenty miles long and 
twenty feet high in the highest part. They are mostly 
inhabited, and the population half a century ago was 
estimated at 50,000, though it has since dwindled down to 


probably a quarter of that number. The presence of so 
large a population must have had some modifying influence 
on the vegetation ; yet not to the extent that might have 
been expected, because there is little cultivation, the natives 
living largely on fish, with which the waters swarm. Mr. 
Woodford says : " The islands are clothed from end to end 
with a dense growth of cocoanut palms and other vegeta- 
tion, and present a beautiful appearance when approaching 
from the sea. The reefs and lagoons teem with fish, thus 
enabling the islands to support a population which for 
their land area was at one time equalled in no part ot the 

Mr. Woodford gives a list of the plants compiled from 
observations on the islands he visited, which he believes is 
nearly complete. As I am able to supplement it by a few 
additional species in the Kew Herbarium, chiefly collected 
by the Rev. Mr. Whitmee, and also to supply specific names 
in some cases where he gives only the generic, I will give a 
list of all the vascular plants known to inhabit the group, as 
a sample of the typical coral island flora. Calopkyllum 
Inophyllum (Guttiferae), Sidafallax (Malvaceae), Triiunfetta 
procumbens (Tiliaceae), Tribulus cistoides (Zygophyllacese), 
Pemphis acidula (Lytheraceae), Rhizophora mubronata 
(Rhizophoraceae), Guettarda speciosa and Morinda citrifolia 
(Rubiaceae), Sccevola Kcenigii (Goodeniaceae), Tournefortia 
argentea (Boraginaceae), Pisonia biennis and Boerhaavia 
^fksYZ (Nyctaginaceae), Euphorbia Atoto? (Euphorbiaceae), 
Ficus tinctoria (Moraceae), Crinum pedunculatum ? (Am- 
aryllidaceae), Cocos nucifera (Palmaceae), Pandanus odora- 
tissimus (Pandanaceae), Fimbristylis glomerata (Cyperaceae), 
Lepturus repens (Gramineae), and Polypodium Phymatodes 
(Filices) — just a score of species, it will be seen, belonging 
to as many different genera, and to eighteen different natural 
'orders of the most diverse habit and structure. They 
are almost without exception plants of general distribution 
in tropical oceanic islands and on the sea-shores of the 
continents. The majority of them indeed inhabit the 
smaller remote islands of the tropical parts of the Indian 
Ocean. I will only add here that their seeds are such as 


are transported by oceanic currents, birds, and winds, with- 
out destroying their vitality. In another article I pro- 
pose discussing these agents of dispersal in some detail. 
The absence from the above list of the two largest natural 
orders — Leguminosae and Compositse — may cause some 
surprise, especially as the seeds of many of the former bear 
long immersion in salt water with impunity, and the pappose 
achenes of the latter are often, it is assumed, conveved 
lont{ distances bv wind. Le^uminos£e are rare in all 
oceanic islands, both coral and volcanic ; but Composite, 
on the other hand, are characteristic of many volcanic 
islands, the Galapagos and St. Helena, for example. 

The distribution of the plants of the Tonga or Friendly 
Islands has been worked out by the writer (5), and a few 
of the most interesting facts may be repeated here. This 
group lies to the south-east of Fiji, between 18 and 23 
south latitude, and 173 and 176° west longitude, and com- 
prises both volcanic and coral islands ; some of the former 
being considerably larger than those of the Gilbert Group, 
and rise to altitudes of 500 to 3000 feet. Fuller informa- 
tion on the geology of the islands will be found in an article 
(6) by Mr. J. J. Lister. But although the Tonga Islands 
are considerably larger than the Gilbert Islands, it is more 
in land area and altitude than external dimensions, and 
it is due partly to the absence of central lagoons. Ton- 
gatabu in the south, the largest of the group, is about 
twenty-two miles in its greatest length, and is composed 
entirely of coral limestone. This island is the best known 
botanically ; but Mr. ]. J. Lister, whose collections were 
worked out for my paper referred to above, thoroughly ex- 
plored the neighbouring smaller, though more elevated, 
Eua, which gave a considerable number of additional 
species. Since the publication of my paper, Kew has 
acquired a collection of dried plants made by Mr. C. S. 
Crosby in the Vavau cluster in the north. This collection 
has not yet been thoroughly worked out, but although 
it doubtless contains some additions, they will not be of a 
character to modify what has been written respecting the 
affinities of the flora of the whole group. The total num- 


ber of assumed indigenous species of vascular plants in 
my enumeration is 290, whereof 246 have a westward, and 
220 have an eastward extension in Polynesia; 138 are 
Australasian (Australia, New Zealand and outlying islands), 
162 are Malayan, and at least 150 have a wider range 
either in the Old or New World, or in both. From the 
foregoing figures it will be seen that the Bora of the 
Tonga Islands is largely composed, like the very small one 
of the Gilbert Islands, of species of wide distribution. 
Indeed no genus is peculiar to the group, and only ten 
species so far as our present knowledge goes are endemic, 
and a more complete exploration of the Fiji Islands and 
other neighbouring groups may reduce this number. The 
290 species of the Tongan flora represent no fewer than 
202 genera and seventy-nine natural orders out of the 202 
recognised in Bentham and Hooker's Genera Plantarum. 
The proportions are 2*55 genera to an order, and 1 '43 species 
to a genus in the Tongan flora. In the flora of the world 
the proportions I obtained by a very rough calculation are 
37 '5° genera to an order, and 12*65 species to a genus. 
Taking the number of Tongan species (138) which extend 
to Australasia, one might overestimate the affinities, be- 
cause, as a matter of fact, a large proportion of these species 
have a wide range. Indeed only a dozen species have 
decidedly Australasian connections. These are : Melicytus 
ramiflorus, Ratonia stipitata, Metrosideros polymorpha, 
Jasmirmm simplicifolium, Hoya australis, Iponuea congesta, 
Pisonia inermis, Peperomia leptostackya, Euphorbia Spar- 
mannii, Ficus aspera, Podocarpus elata and Pteris comans. 
It will be perceived that the connections are specific rather 
than generic. But the most significant facts brought out 
in the paper under consideration are two, namely, the 
large proportion of species — upwards of a third — peculiar to 
Polynesia, and the strongly Malayan character of the flora, 
generally, of the Tonga, Fiji and Samoa Islands. 

Several additional small contributions to the flora of 
the Solomon Islands have appeared (7), including some 
highly interesting novelties collected by the officers of 
H.M.S. Penguin, and the Rev. R. B. Comins. Excellent 


photographs of the singular new genus Sararanga 
(Pandanaceae) have been received at Kew, as well as ripe 
fruit in spirit, which will enable me to add to my published 
description, though not to complete it, because the male 
inflorescence is still unknown. Two species of Begonia, an 
Oxymitra (Anonacese) with flowers nearly nine inches long, 
a singular Tabernce Montana having a twisted fruit, and 
the anomalous genus Lophopyxis (8) are among the latest 
additions to the flora of the Solomon Islands. The last is 
doubtingly placed in the Euphorbiaceae by Sir Joseph 
Hooker, and it has since been twice described (9 and 10), 
and placed in different natural orders, namely, Combretopsis 
(Olacinese) and Treubia (Saxifragaceae). There are two 
or three very closely allied species or races inhabiting 
Malacca, Ceram, New Guinea, and the Solomon Islands. 
I may refer in passing to a zoological paper (11) in which 
the author puts forward the theory of a former connection 
of the Solomon, Fiji, New Hebrides, Loyalty, New- 
Caledonia, Norfolk and New Zealand Islands with New 
Guinea, but not with Australia. That there was, in the 
remote past, a greater land area in this region seems 
highly probable, but the relationships are so complex that 
fuller data are required to afford a solution of the problem. 
The present flora of Lord Howe Island, described a few 
pages forward, does not favour Mr. Hedley's views in their 
entirety on this point. 

In my reference to the flora of Christmas Island (12) I 
overlooked a paper that supplemented mine to some extent 
(13), especially in relation to the vegetation. 

Dr. Trimen (14) has published two more volumes of 
his admirable flora of Ceylon, bringing it down to the end 
of the Balanophoraceae, following the arrangement of 
Bentham and Hooker's Genera Plantarum. The same 
author has drawn up a provisional list (15) of Maldive 
plants ; the first, I believe, that has appeared. As might 
be expected there is no endemic element, and the vegeta- 
tion is an assemblage of the ubiquitous coral island plants 
and weeds of cultivation. Dr. Trimen makes no mention of 
the Cocos maldivica or Coco-de-mer (Lodoicea sey die liar inn) \ 


but, although it is improbable that this palm ever grew in 
the Maldive Islands, something yet remains to be done to 
complete its history. John de Barros, a Portuguese 
author, is thus quoted (16) by the writer of an article on 
these islands : — 

"Their productions he also enumerates minutely, especially 
the coconut, both of the ordinary kind and of that called 
coco-de-mer, almost peculiar to the Seychelles, the seed of 
which appears to have been borne thence to the Maldivas 
by the currents of the ocean ". 

Since the publication of my notes on the flora of New 
Zealand and the outlying islands (17) several interesting 
papers on the subject have appeared, though there is only 
one of sufficient importance to call for more than brief 
mention. But first the minor ones. Mr. F. Kirk is the 
author (18) of a series of monographs treating of the 
genera Gentiaua, Colobatttktis, and Gunnera, as re- 
presented in the New Zealand region, besides descriptions 
of a number of new species belonging to various natural 
orders. The forms of Gentiana are numerous, and the 
species exceedingly difficult of delimitation. Kirk defines 
ten species, and about half of them comprise several 
varieties. They are spread all over New Zealand, except 
the extreme north, and they extend to the Chatham, 
Antipodes, Auckland and Campbell Islands ; but hitherto 
no species has been found in Macquarie Island, the southern- 
most of these islands. They chiefly inhabit the mountains, 
in alpine and subalpine situations, and the sea-coast ; four 
out of the ten, it is stated, not being found out of the reach 
of the sea-spray. They all belong to one group, char- 
acterised by having pentamerous flowers, unappendaged 
corollas, and versatile anthers. White is the prevailing 
colour of all the species, though some of them occasionally 
exhibit various shades, mostly dull, of red, purple, and violet, 
and more rarely a pale yellow. This is in direct contrast to 
the behaviour of the northern species, speaking generally, and 
we are indebted to Mr. Kirk for the observation. Colo- 
banthus (Caryophyllaceae) is one of those densely tufted 
moss-like genera of which there are representatives in 


various natural orders. It is one of the very few genera 
common to Australasia, to the Antarctic, and other southern 
islands, and the Andes, and confined to these regions. One 
species, C. quitensis, ranges from the mountains of Mexico 
to Cape Horn and reappears in New Zealand. Kirk also 
records it from Amsterdam Island, but that seems to in- 
volve two errors, for, so far as our data at Kew go, C. 
diffusus inhabits St. Paul, and no species is found in the 
neighbouring island of Amsterdam. One species, C. 
Billardieri, is found in the Alps of Victoria, in Tasmania, New 
Zealand, and the small islands southward to Macquarie. Two 
Falkland Islands species also recur in South Georgia, the 
southern insular limit of phanerogamic vegetation in the Pata- 
gonian region, if we except a grass, Aira antarctica, collected 
by Dr. Eights in the South Shetlands, about 62 S. lat., or 
8° south of South Georgia. Kirk enumerates and de- 
scribes ten species of Colobanthas from the New Zealand 
region, including four proposed new ones. 

Gunnera (Haloragidacea^) has a similar range to that 
of Colobantkus, save that it does not reach the colder limits 
either in America or the New Zealand region. Kirk 
brings up the species of the latter region to nine, four of 
which are new. 

W. Colenso, D. Petrie, and H. C. Field also describe 
a few novelties (19), and the first named gives a charming 
description of his travels and botanising in the romantic 
country around Hawke's Bay, upwards of fifty years 

The one paper which I propose to discuss a little more 
in detail is devoted to the natural history of Macquarie 
Island (20), the most southerly speck of land in the New Zea- 
land region known to support phanerogamic vegetation. It 
is in the same latitude (54 S.) as South Georgia in American 
waters, the flora of which I have described (21), where a 
list is given of the vascular plants inhabiting the island. 
They are separated from each other by about 164° of 
longitude, which in this latitude means, in round numbers, 
5875 geographical miles ; yet, as previously stated, nine 
out of thirteen of the vascular plants found in South 


Georgia also occur in some of the southern islands in the 
New Zealand region. Later on I shall have something- 
to say, or rather repeat, in explanation of this fact. It 
should be noted that these islands are in about the same 
latitude as York in England ; yet the climate is now so 
severe in South Georgia and other conditions are so un- 
favourable to vegetation that the flora is perhaps poorer 
than in the highest northern latitudes yet explored, and 
entirely wanting the colour characteristic of many northern 
flowers. For example, such charmingly beautiful plants as 
Papaver nudicaule, Silene acaulis, Saxifraga oppositifolia 
and Epilobium latifolium are found north of the eightieth 
parallel ; whereas the showiest flowers in South Georgia are 
those of a very small buttercup, so small indeed that they want 
finding. The flora of Macquarie Island is, however, not 
altogether devoid of colour, as witness Pleurophyllum ; and 
Stilbocarpa is remarkable for its large rhubarb-like leaves. 

Macquarie Island is between twenty and twenty-five 
miles long and five or six miles across in its broadest part. 
It is generally hilly, though the hills are nowhere above 
800 feet. The following is a list of the vascular plants 
recorded by Mr. Hamilton (20), who visited the island 
early in 1894. I may mention that I had most of these 
plants under observation (22), and I do not agree in every 
instance with his and Mr. Kirk's (23) determinations ; but 
the divergencies are unimportant ; and there are several 
corrections of the names given in previously published lists. 
Ranunculus ci-assipes, Cardaminc hirsuta, var. corymbosa, 
Colobanthus muscoides, C. Billardieri, Stellaria decipiens, 
Mont ia font ana, Aceena Sanguisorbcz, A. adscendens, Calli- 
triche a7itarctica. Epilobium nummularifolium, E. lin- 
nceoides, Azorella Selago, Stilbocarpa polaris, Coprosma 
repens, Cotula plumosa, Pleurophyllum Hookerii, Uncinia 
nervosa, Luztila criuita, Deschampsia Hookeri, D. penicil- 
lata, Poa foliosa, P. Hamiltonii, Agrostis antarctica, 
Festuca contracta, Aspidium aculeatum, var. vest it um Poly- 
podium aust?'ale, Lomaria alpina and Lycopodium Billar- 
dieri, var. varium. The last named one would have 
hardly expected to find in so high a latitude, where the 


only woody plant is the small creeping Coprosma repens, 
because it usually grows on trees. A re-examination 
of a very small collection of Macquarie Island plants 
sent by Mr. Fraser of the Sydney Botanic Garden to the 
late Sir William Hooker, about sixty years ago, has led to 
the discovery of Lycopodium Selago, associated with Azor- 
ella Selago, a very similar plant in external appearance. In 
addition to the foregoing there are three colonised vascular 
plants, namely, Stellaria media, Cerastium triviale and Poa 
annua ; and Mr. Hamilton states that he also collected 
Tillcea muscosa and two sedges, but the specimens were 
lost. If we except three imperfectly known grasses, which 
Mr. Kirk has described as new (24), there are no endemic 
plants in the island. The vascular cryptogams are all 
widely spread, two of them recurring in the northern hemi- 
sphere. Of the flowering plants upwards of half are confined to 
the New Zealand region, and the rest have a wider range. 
Stilbocarpa polaris ( Aral iacese) and Pleurophyllum Hookerii 
(Composite) are the two most remarkable and most con- 
spicuous plants in this meagre flora ; the former having 
large rhubarb-like leaves, and the latter silky, silvery leaves 
and handsome purple flower-heads in long racemes. Colo- 
banthus, Azorella, Acczna and Uncinia are equally charac- 
teristic in the South American region. 

Quite recently a fresh account of Lord Howe, Pit- 
cairn and Norfolk Islands has appeared (25), but it con- 
tains nothing new on the botany of these islands. Special 
stress is laid on the beauty of the vegetation of Howe 
Island, where palms and tree ferns abound, and fig-trees of 
the banyan type attain dimensions hardly exceeded else- 
where. What is known, however, of the botany of this 
interesting island has appeared in Government Reports and 
scattered in a variety of publications (26-29) of limited 
circulation. It is true that Sir F. von Mueller long ago 
published (30) a bare list of all the plants known to him 
from the island, but it is incomplete, and supplies no in- 
formation beyond the names of the plants. This being so, 
I am preparing a detailed account of the flora of this island 
with a view to publication elsewhere. I may here give, 


however, some particulars gleaned from the publications 
referred to, though they are mostly anterior to the date 
(1885) to which I have limited myself generally in these 
articles, adding a few remarks of my own on the distribu- 
tion of the plants. 

Lord Howe Island is of small extent and peculiar con- 
formation, situated about 300 miles from the coast of 
New South Wales in 31 35' S. lat. It is seven miles 
long with an average breadth of one mile, and the steep 
circular flat-topped elevations rise to a height of nearly 
3000 feet. Norfolk Island, the nearest land to the north- 
east, is about 500 miles distant, and New Zealand, to the 
south-east, somewhat farther off. The island is of volcanic 
origin, consisting of three basaltic masses connected by 
coral-sand rock. About 165 species of indigenous flower- 
ing plants are known, and forty-eight ferns and lycopods. 
As already indicated palms form a conspicuous feature in 
the scenery. There are four species, all endemic, and 
they have been very much named, though three out of 
the four are well known under the generic name of Kentia. 
They are K. Belmoreana, K. Canterburyana and K. For- 
steriana — names familiar to many persons, as they have 
long been favourite palms in cultivation on account of their 
elegance and hardiness. A tall and graceful specimen of 
K. Forsteriana is one of the finest ornaments of the central 
part of the palm-house at Kew. The fact of there being 
a good market for the seeds of these insular palms has led 
to considerable destruction of the trees to obtain them ; but 
I believe the Government of New South Wales has made 
it a punishable offence to destroy trees on public territory. 
Beccari (31) has founded the genus Howea for them, which, 
if accepted, is the only endemic one. There are also four 
indigenous tree ferns, three of which are endemic. But the 
banyan trees {Fiats columnaris) are perhaps the most 
striking objects in the vegetation. Several appear in the 
photographs illustrating Wilson's Report, one of which is 
said to cover an area of three acres ! Morcea Robinsoniana 
is an outlying gigantic member of an African genus of 

Irideae very closely allied to Iris itself. It is known as the 



wedding-flower, and there is a fine specimen of it at the 
south end of the cactus-house at Kew. Carmichcelia exul 
(Leguminosae) is the only species of a considerable genus, 
with this exception, not known to inhabit any other country 
than New Zealand. There are other connections with the 
flora of the latter country, but they are mostly such as extend 
to Australia as well. Pimelea longifolia and the handsome 
sedge, Gahnia xantkocarpa, are apparently exceptions. In 
round numbers 25 per cent, of the species of flowering plants 
of Lord Howe Island are endemic, and 62 per cent, are 
common to Australia, many of these having a wider 
range. A few are common only to Australia, New Zealand, 
and Norfolk Island. The shrubby violaceous genus 
Hymenanthera is an example. The gum trees {Eucalyptus) 
of Australia are represented by the endemic Acicalyptus 
Fullagari, a small Fijian genus differing from Eucalyptus 'in 
having a calyptrate calyx-limb and separate petals. Two 
other conspicuous trees in the endemic element are Draco- 
phyllum Fitzgeraldii (Epacridese) and the screw-pine, Pan- 
danus Forsteri. The former is a tree, said to be the largest 
in the order, attaining the height of fifty to sixty feet. It has 
the foliage and aspect of a monocotyledon rather than of a 
dicotyledon. One characteristic Australasian type we miss 
in the Lord Howe Island flora, and that is Cor dy line. 

When reviewing (32) the newer literature relating to 
the flora of the Galapagos Islands I found little to add to 
what had been done by Darwin, Hooker and Andersson ; 
merely mentioning the visit of the United States ship 
Albatross, and Dr. G. Baur's theory of the origin of the 
fauna and flora. Since then an account of Dr. Baur's 
botanical collections has been published (33), and the sub- 
stance has also appeared in an English journal (34), and 
Dr. Baur himself has written (35) and lectured (36) in 
defence of his theory of the origin of this group of islands. 
As previously stated, he contends that the evidence points 
to the present condition being the result of subsidence ; 
that the islands were formerly connected with each other 
and at a still earlier period with continental America. 
Although this theory has been derided, I think the biologi- 


cal data strongly favour its correctness, and the soundings 
given in the map accompanying Agassiz's report (2,7) of 
the Albatross expedition show a relatively shallow area 
in which the Galapagos Islands are situated, and which 
extends eastward to the mainland of Veraguas. Probably 
the separation would be greatly anterior to the segregation 
of the West Indian Islands. 

In the Botany of the Challenger expedition (38) I 
attempted a rough classification of islands in relation to the 
composition of their floras. These are defined as follows : 
1, Vegetation comprising a large endemic element including 
distinct generic types ; 2, vegetation comprising a small, 
chiefly endemic element, the derivation of which is easily 
traced ; and 3, vegetation containing no endemic element. 
Without due consideration the Galapagos were referred to 
the first category. Sir Joseph Hooker (39) fully realised 
the absolute American affinities of the flora ; but he analysed 
and discussed it as a derived one rather than as a remnant. 
Darwin, through some misinterpretation of the statistics sup- 
plied to him, fell into a singular error respecting the generic 
endemic element in the Galapagos (40). Referring to 
the Compositse, he says : " There are twenty-one species, 
of which twenty are peculiar to this archipelago ; these 
belong to twelve genera, and of these genera no less than 
ten are confined to the archipelago ! " How this error arose 
it is impossible to say, but as a matter of fact the statement 
quoted is wrong (and was wrong at the time it was written) 
in all its details. With regard to assumed endemic genera 
of Compositae, five were founded on galapageian plants, 
namely, Microcoecia and Desmocephalum, since reduced to 
Elvira ; Macrcea to Lipochczta ; and Scalesia and Lecocar- 
pus are so near to Mirasolia and Melampodinm respectively 
that the late Mr. Bentham gave it as his opinion that they 
might well be reduced. Two genera from these islands 
belonging to other orders have also been reduced. These 
are Galapogoa = Coldtnia (Boraginacae), and Dictyocalyx 
= Cacabus (Solanacese) ; and Pleuropetalum (Amarantaceas) 
has since been found in several localities in Western 
America. Taking this view of their affinities, there is not 


a single genus of flowering plants endemic in the Galapagos ; 
but each island has its distinct species. Briefly put then, 
the genera are the same in all the islands, and the genera 
are American ; whereas a large proportion of the species 
are peculiar to each island, though they are not so ex- 
clusively confined to single islands as Darwin supposed. 
On this point he says (41) : " Again Euphorbia, a mundane 
or widely distributed genus, has here eight species of which 
seven are confined to the archipelago, and not one found on 
any two islands. Acalypha and Borreria, both mundane 
genera, have respectively six and seven species, neither of 
which genera has the same species on two islands, except 
in the case of one species of Borreria." Dr. Baur's 
recent explorations necessitate a considerable modification 
of this statement ; yet in a sense they confirm and empha- 
sise it. Baur himself deals more particularly with the fauna 
(36) in illustration of this phenomenon. More than 400 
specimens of the lizard genus Tropidurus were collected, 
and in the result he found that "each island possessed only 
a single species ; all the individuals of an island belonged 
to one species ; and nearly every island had its peculiar 
species or race ". 

The botanists who worked out Dr. Baur's collections 
selected Euphorbia viminea (33) as an example of a plant 
exhibiting racial differences in each of the eight islands, 
where it is now known to occur. The genera Acalypha 
and Borreria are cited as other instances. On the other 
hand, Euphorbia articulata, which was collected on four 
different islands, showed no such tendency. 

In a former article in this journal (32) I mentioned the 
fact that huge branching Cactacese form one of the most 
striking features in the lower zone of the vegetation of the 
Galapagos, and I have elsewhere (42) given some par- 
ticulars of what is known, and how little is known of these 
Cactaceae ; and I may repeat here that specimens of only 
one species have, so far as I can ascertain, been brought 
away from the islands. These were brought to this country 
by Darwin, and published by Henslow (43) under the 
name of Opuntia galapageia. This species is remarkable 


in the genus for its very small flowers, which are only about 
three-quarters of an inch in diameter, and also for the small 
number of petals ; but as the figure was made from dried 
specimens, it may be inaccurate in some details. In the 
same place it is mentioned that a species of Cereus was 
common in the island, but was not found in flower. 

Darwin himself specially alludes (44) to the prominent 
feature these Cactacese are in the landscape, and likewise 
to the fact that they grow in the rough lava where there is 
absolutely no other phanerogamic vegetation. He further 
points out their importance as food for the gigantic tor- 
toises and land lizards. They are also a source of water 
during the severe droughts, which often parch the lower 

Subsequent travellers have dwelt upon the part the 
Cactacese play in the biology of the island, and Andersson, 
a botanist who visited the islands in 1852, states (45) that 
he observed four or five species, but had time neither to 
prepare specimens nor sketch the plants. 

My note on the subject in Nature came under Dr. 
Baur's notice, and he forwarded me two photographs, one re- 
presenting a fine example of an arboreous Opuntia of great 
size, and the other a view embracing a number of large 
Cerei, together with a transcript of his notes on the subject 
in a paper (46) which I had not seen. He was struck by 
the difference in the appearance of the Optmtice on the 
different islands, and observed that the large Opuntia has 
a different habit on nearly every island. Thus, on Barring- 
ton, Indefatigable and South Albemarle, it develops a 
very tall stem ; on Charles and Hood a relatively short 
but thicker stem ; on Jervis a very short stem, branch- 
ing from very near the ground, and on Tower Island 
it forms no stem at all, and appears as a dwarf bush. 
Dr. Baur attributes these modifications to the varying degrees 
of humidity, the greatest development occurring in the driest 
climate. In the lower region of South Albemarle, up to 
about 500 feet, the Opuntia is very common, attaining a 
large size, the largest being about twenty feet high, with a 
trunk two feet in diameter. " In old trees the bark looks 


very much like that of a pine, and peels off in very thin 

The common Cereus, which strongly resembles C. 
peruvianus, attains almost the same dimensions ; but this 
is all we know about it at present, and there is clearly 
much more botanical work to be done in the Galapagos 
before the subject is exhausted. It may be of interest to 
add that no species of cactus inhabits the island of Juan 
Fernandez, but this may be ascribed to climatic differences. 
Indeed, so far as is known, none of the other Pacific American 
islands, at any considerable distance from the coast, support 
any members of the order, though Malpelo, for example, 
is barren enough to give them a chance of flourishing. 

Another remarkable element in the flora of the Galapagos 
is the relatively large number of species of the small order 
Amarantacese. About fifteen species are now known to in- 
habit the islands, and twelve of them are endemic. They 
belong mainly to the genera Telautkera, Alternanthera, 
and Froelichia. 

Concerning the flora of the Arctic Islands in relation to the 
adjacent continents, I have to add a few references (47-48) 
to works of older date than my paper (49), and a few recent 
ones of unusual interest. Mr. Trevor- Battye's account of 
the vegetation of Kolguev Island (50) and Colonel Feilden's 
contributions on the subject (51-52) rank first among these. 
The former noted ninety-five species of phanerogamia in 
Kolguev, and his observations on the vegetation are of 
great value. About a score of the plants recorded by 
Ruprecht (53) were not found, and Trevor-Battye remarks 
on the absence of Saxifraga oppositifolia, Mertensia maritima 
and Ledum palustre. Colonel Feilden's short paper on 
Spitsbergen plants, as well as his remarks on mild arctic 
climates, is worthy of attention on account of his experience. 
The only information I have found (54) respecting the 
vegetation of Einsamkeit Island is that there is no grass 
carpet, and it is added that there is a great quantity of drift- 
wood, sometimes far inland. A new list (55) of Iceland 
and Faeroe plants does not claim to be anything more than 
a contribution to local distribution. 


There is little new literature relating to the Atlantic 
Islands, but Sir Joseph Hooker's comparison (56) 
of the Maroccan and Canarian floras was overlooked by me 
when reviewing the writings of Dr. Christ. In an article 
(57) of more recent publication, the latter gives expression 
to a considerable modification of his views on the affinities 
of the Canarian flora. He now recognises a much more 
intimate connection with the old African flora. But I must 
not reopen the subject here. 

One important contribution (58) to the flora of the West 
Indies has appeared. This part consists of a critical 
elaboration of the Myrtaceae, than which there was probably 
no group of plants more in need of revision. It is some- 
what appalling to see such familiar trees as the allspice and 
clove with a page and half of synonyms each ; yet it is 
very useful, historically, as well as for practical purposes, to 
have them brought together. 


(To be continued.} 



THUS far I have tried to rehabilitate the cell as a vital 
unit. I have now to deal with the further question as 
to the part played by the cell in the composition of the higher 
animals and plants. In the earlier part of this essay I 
stated that Mr. Adam Sedgwick denied in toto the proposi- 
tion that "the elementary parts of all tissues are composed 
of cells ". Since writing those words, Mr. Sedgwick's reply 
to my previously published criticisms has appeared, 1 and I 
find that I have made a mistake. For he does not deny 
the proposition, but says: ''The assertion that organisms 
present a constitution which may be described as cellular is 
not a theory at all ; it is — having first agreed as to the 
meaning and use of the word cell — a statement of fact and 
no more a theory than is the assertion that sunlight is com- 
posed of all the colours of the spectrum ". I can only beg 
Mr. Sedgwick's pardon. I certainly was led to suppose 
from his earlier writings that he regards the cell as a 
nonentity, in so far as it may be considered to be the 
ultimate structural unit of the metazoa, and I recoiled from 
his suggestion that the essence of development lay in "a multi- 
plication of nuclei and a specialisation of tracts and vacuoles 
in a continuous mass of vacuolated protoplasm ". 

Mr. Sedgwick now explains that he objects, not to the 
statement that tissues are composed of cells — or, in his own 
words, that they have a composition which may be described 
as cellular — but to the statement that an individual meta- 
zoon is an aggregate of lesser individuals, or, as it has often 
been expressed, a cell colony or cell republic. I have else- 
where — and as Mr. Sedgwick well says, after great effort — 
come to agree with him on this point, for a careful survey 
of a considerable range of facts led me to the conviction 

1 Adam Sedgwick, "Further Remarks on the Cell-Theory, with a Reply 
to Mr. Bourne," Quart, four. Micr. Sci., vol. xxxviii., p. 331, 1895. 


that the idea of a cell republic was inappropriate. Such 
being the case I would willingly have buried the hatchet, 
but when I had already dug the hole to bury it in, my hand 
was stayed by some criticisms on his views and on mine 
which have just been published in a contemporary periodi- 
cal. 1 These criticisms have restored to me the conviction 
which I held when I ventured to write a criticism of Mr. 
Sedgwick's views ; a conviction that, as he originally 
expressed them, they were calculated to mislead and to do 
harm to the very cause whose interests he was desirous to 
promote. As he has lately explained that he did not mean 
what I supposed him to mean, there is no need for quarrel- 
ling any further with him, but he will himself allow that I 
was amply justified when I gave the following as a not 
unfair statement of his position. That from the connection 
known to exist between some cells composing adult tissues, 
there is an antecedent probability that similar connections 
exist between all cells composing all tissues ; and this 
probability is heightened by observations made on the 
development of Peripatus, by the fact that the so-called 
mesenchyme cells in Avian and Selachian embryoes are 
continuous and not isolated as was once supposed, and by 
a study of the developing nerves of Elasmobranchs. And 
that it follows from this that the morphological concept of a 
cell so far from being of primary is altogether of secondary 
importance, and that progress in the knowledge of structure 
is impossible so long as men persistently regard cells as the 
fundamental structural units on which the phenomena mani- 
fested by organised beings depend. The true method of 
inquiry must be a study of the growth, extension, vacuolation 
and specialisation of the living substance protoplasm. 

He has been understood by others as I understood him, 
and indeed he had so expressed himself that he could 
scarcely have been understood otherwise. What I had 
anticipated has happened. Persons, ready to grasp at 
novel ideas, have said in their hearts : "Tush, there is no cell ! 
There are protoplasmic masses which may contain one or 
many nuclei ; the mass is of no importance, it is scarcely 

x Natural Science, vol. vii., No. 46, December, 1S95. 


more than the medium in which the nucleus lives, and 
through which it exhibits its powers. The nucleus may 
move about in the mass, acquiring ' spheres of influence ' 
at its halting places, and so producing the vital phenomena. 
It is the nucleus which is the vital unit, and there is no 
bond between nucleus and cytoplasm which shall compel 
us to regard their union as the necessary condition of living 

I have made use of my own expressions, but if this is not 
the plain meaning of the short editorial entitled "The 
Reign of the Nucleus" in the January number of Natural 
Science, what is ? 

The writer of the editorial is so captivated with the pros- 
pect opened up by his interpretation — a perfectly legitimate 
interpretation — of Mr. Sedgwick's writings, that he forthwith 
abolishes the existence of cells altogether and talks glibly 
of " protoplasmic masses," ignoring the fact that the masses 
in question are divided up into corpuscles. Following up 
his theme of protoplasmic masses dominated by nuclei, he 
lightly dismisses the arguments which I put forward, 
saying that the segmentations of Nereis, Unio, etc., exhibit 
nuclear lineage rather than cell lineage (who could 
hold such an opinion after a careful study of Wilson and 
Lillie's figures ?), and winds up with the following astonish- 
ing piece of criticism : " In drawing an argument for the 
cell-theory from the definite places assigned to cells in 
development Bourne seems to us to have overlooked the 
experiments of Wilson, Driesch and Hertwig, who have 
shown that the nuclei may be moved about in the proto- 
plasmic mass almost as freely as a ' heap of billiard balls 
may roll over each other ' ". I rubbed my eyes and 
wondered. I thought I knew the works of Driesch, 
Hertwig and Wilson pretty well, and that I had considered 
them carefully, and I had certainly regarded them as strong 
evidence in favour of the cell-theory as I conceived of it. 
A short search soon hit upon the passages which are 
professedly quoted. First for Driesch : x " Die Furchungs- 

1 Hans Driesch, Entwicklungmechanischse Studien, iv., Zeitschrift fUr 
IViss. Zoologie, vol. Iv., 1893. 


kugeln der Echiniden als ein gleichartiges Material 
anzusehen sind, welches Man in beliebiger Weise, wie 
einen Haufen Kugeln durch einander werfen kann, ohne 
dass seine normale Entwicklungsfahigkeit darunter im 
Mindesten leidet ". (The segmentation spheres of Echinids 
are to be regarded as a homogeneous material which one may 
roll amongst one another at will like a heap of balls, without 
thereby destroying in the least their capacity for develop- 
ment.) No hint whatever of rolling the nuclei through the 
protoplasmic mass. The statement is made of Furchungs- 
kugeln, that is of cells, and it is the cells that one may roll 
about like balls. Not a bad argument for my contention, 
that the blastomeres of many developing ova are disjunct. 
If there were any doubt as to Driesch's words a study of 
figures 39-68 which illustrate his paper would satisfy the 
most exacting. The blastomeres are unusually distinct 
from one another, especially in the embryoes illustrated by 
figs. 63 and 67. Now for Hertwig: 1 "Bei den verschiedenen 
Modificationen des Furchungsplasma werden die aus dem 
ersten Furchungskern durch aufeinanderfolgendeTheilungen 
erzeugten Kerngenerationen Theilen des Dotters, die in Eir- 
aum eine sehr verschiedene Lage einnehmen, zuoetheilt und 
mit ihnem zu einem zellkorper verbunden. Die Kerne wer- 
den in Eiraum wie ein Haufen von Kugeln durch einander 
gewurfelt." This is a very complicated German sentence 
and might well lead to a misunderstanding, but it comes out 
all right in plain English. "In the various modifications of 
the divisional processes the nuclear generations, which are 
produced by successive divisions from the segmentation 
nucleus, are assigned to a portion of the yolk which occupies 
very different positions within the limits of the egg, and are 
bound with it to form a cell body. The nuclei are rolled 
one over another within the limits of the egg like a heap of 
balls." This passage is a summary of preceding state- 
ments and inferences, and it might be held to bear a very 
different meaning to that which it does bear ; the illustration 

1 O. Hertwig, " Ueber den Werth der ersten Furchungszellen fur die 
Organbildung der Embryo," Arch, fur Mikr. A nat., vol. xlii., p. 662, 1893. 


of the heap of balls is a very loose one. To understand 
the meaning of the summary one must turn to pp. 678-685 
of the same memoir, which consist of a section entitled 
" Erklarung des abnormen Furchungsverlaufes ". There 
we learn, as we had previously learnt from Driesch, that 
the divisional planes of segmenting ova are determined by 
the direction of the nuclear spindles and that the orientation 
of the first nuclear spindle is determined by the character of 
the body of the ovum and its contents. The ova of Echinus 
are homogeneous throughout, and orientation of the first 
nuclear spindle is a chance affair. But the ovum of the 
Frog is not homogeneous ; it consists of a smaller cap of 
protoplasm resting on a large body of yolk, and the nucleus 
lying in the cap of protoplasm, the direction of the first 
nuclear spindle is determined by its relations to the more 
active yolk on the one hand, and the denser food yolk on 
the other. The relations of the food yolk and protoplasm 
are changed by the pressure applied during the experiments 
and the changes are different according as the pressure is 
applied vertically or horizontally. Hence the direction of 
the first and the succeeding nuclear spindles is changed 
in different senses, according to the pressure employed. As 
the divisional planes are always at right angles to the 
nuclear spindles, the positions of the two first and the suc- 
ceeding blastomeres differ according as the pressure applied 
is vertical, horizontal, oblique, or circumferential. One may 
in fact cause the blastomeres and their contained nuclei to 
take up what position one will by varying the direction of 
the pressure. In this sense, and in this sense only, can one 
speak of rolling the nuclei about like balls. Not a word about 
a protoplasmic mass through which the nuclei are caused to 
roll. On the contrary, a great deal about planes of division 
and splitting up of the egg into corpuscles round the nuclei. 
It only requires a glance at Hertwig's figures and diagrams 
to show that the blastomeres are as distinct during abnor- 
mal division as during normal division, and that there is not 
at any time any question of a "protoplasmic mass," a cir- 
cumstance which has been well understood by everybody 
who has taken the trouble to read his memoir carefully. 


Most of the experiments of Wilson, Hertwig and 
Driesch were of a different kind. They isolated the blas- 
tomeres by gentle shaking. Driesch is very careful to say 
gentle ; rough shaking destroyed the individual blastomeres. 
Things which are so loosely united as to be separated thus 
easily from one another scarcely suggest the nature of a 
coherent protoplasmic mass. 

The criticism falls entirely to the ground and one can 
only wonder how any one could have had the temerity to 
make it. The very objections urged to my views are but 
additional evidence in support of them, and I was well 
aware that the evidence existed when I wrote, but I had to 
be as brief as possible, and did not refer to it. My state- 
ment that it is very clearly established that there are 
numerous cases in which there is not "a primitive con- 
tinuity which has never been broken" is abundantly 
justified. Mr. Sedgwick wonders why I emphasised the 
distinction and complete isolation of the cells formed by 
the segmentation of the egg. The reason is surely clear 
enough. Because he suggested, in no uncertain manner in 
his earlier writings, that the connections between adult cells 
were due to a primitive continuity which had never been 
broken, and that those who urged that such connections 
were secondary were in the wrong. This suggestion was 
contrary to fact, and it was my object to show that it was. 
I did not contradict myself when I stated immediately 
afterwards that the organism cannot be considered to consist 
of independent life units, for I went on to show that the 
cell-republic theory is also contrary to fact, and must there- 
fore be condemned. If a contradiction exists, it exists in 
nature, and after we have ascertained the facts the next 
thing is to try to explain this seeming contradiction. Mr. 
Sedgwick says that he does not think it possible to do so, 
until we acquire some more understanding of the relative 
functions of nuclei and protoplasm. Possibly he is right, 
yet I think that an attempt may be made, and if the explana- 
tion is after all not very satisfactory yet some service may 
be done, for we may arrive at more distinct ideas about 
fundamental points, and we must gain much by a careful 


classification of the facts. Such a classification has yet to 
be made. So long as a theory is dominant, as the cell- 
republic theory was, exceptions and difficulties are glossed 
over, or are explained away by a phrase. When I made a 
vigorous onslaught on Mr. Sedgwick, I was afraid that he 
wished to substitute King Stork for King Log and bring us 
under the domination of a new theory of his own. His 
reply to my strictures and his careful exposition of his own 
standpoint are reassuring on this point, and if I exceeded the 
limits of courtesy in my article, I did so under a misunder- 
standing and express my regret for it. Mr. Sedgwick has 
done a great service in breaking the bonds of the old theory. 
Now the question is, having got our liberty, what are we 
going to do with it ? 

Firstly, I think, we have got to make up our minds as 
to what we mean by a vital unit. 

In the first part of this essay I stated that the cell is par 
excellence the vital unit, by which I meant nothing more 
than that it is the simplest form of material aggregate in 
which individual life is possible. There would seem to be 
no objection to such an application of the word unit. But 
the term unit is a relative one, and its correlative is 
multiple. If, therefore, we see that the developing embryoes 
of many animals and likewise the tissues of the adult forms 
are made up of structures which we must call cells, and if 
we call the cell a vital unit, we are obliged to conclude that 
the animals in question are composed of an aggregate of 
vital units, which leads us directly to the doctrine of a cell- 
republic. Thus at the outset we are confronted by the 
great difficulty that what experience teaches us to deny 
reason compels us to affirm. 

There must be a flaw somewhere, either in the facts or 
in the reasoning. There can hardly be any doubt about 
the facts ; the flaw therefore must be in the reasoning, and 
I do not doubt that it consists in our insistence on applying 
the idea of a unit to biological facts. As Whewell would 
have said, the idea is inappropriate. The term unit, as we 
use it in Biology, conveys a double meaning. On the one 
hand, it borrows part of its meaning from the idea of num- 


ber, and to this extent the term is used in an equivalent 
sense to that in which it is used in Physics. But put side 
by side such expressions as unit of mass or unit of time with 
the expression unit of life, and a little reflection will suffice 
to show that the sense is inappropriate. Nor is the case 
made better if we compare the unit of life with the chemical 
unit. The value of the latter consists essentially in this, 
that it is a means of dealing numerically with chemical facts, 
and experience shows that ideas of number are very 
appropriate to chemical facts. With life the case is very 
different. In the present state of our knowledge the con- 
nection between life and number is of the slenderest kind, 
and it is insufficient to justify our applying numerical ideas 
to vital phenomena. 

The other sense in which the term unit is used in 
Biology is purely subjective. It stands to express our idea 
of individuality, an idea which is founded on our own states 
of consciousness. It is unnecessary for me to dilate upon 
the controversies which have raged round this idea of in- 
dividuality in its application to the animal kingdom. The 
most acute reasoners are not agreed upon the precise 
point where individuality ceases to belong to parts and 
belongs to the whole even in some of the simpler colonial 
organisms, and in such cases as the Siphonophora a satis- 
factory solution of the problem appears to be hopeless. 

But these cases are simple in comparison with that 
which we are now discussing. If then we cannot agree 
about the limit of individuality in colonial organisms, how 
are we likely to agree about the same thing in the case of 
organic structure in general ? 

There is this to be said, however, that for us the test of 
individuality should be a biological test, and the idea is there- 
fore more appropriate to the question than the numerical idea 
just spoken of. It was, no doubt, the recognition of its 
propriety which lent such force to Schwann's argument, 
" since it may be proved that some cells, which do not 
differ from the rest in their mode of growth, are developed 
independently, we must ascribe to all cells an independent 
vitality ". 


Hence, as it seems to me, whilst we can and ought to 
get rid of the numerical idea expressed by the word unit, we 
cannot get altogether rid of the idea of individuality, and 
we must do our best to bring it into harmony with the facts. 

Since there is an inseparable connection between the 
idea of number and the word unit, we ought to get rid of 
the expression " unit of life," and use some other term 
which shall denote alike the simplest and the most com- 
plex of living beings. The word organism I have aheady 
objected to because of its double connotation — would it not 
be better to make use of such a word as "biont," which is 
as nearly as possible the equivalent of the German " Leben- 
diges " ? Anything which leads or is capable of leading an 
independent individual life is a biont. Thus a cell may be 
a biont, as in the case of the protozoa, or it may be a con- 
stituent part of a biont, as in the case of the metazoa. In 
any case the cell is the simplest form of biont known, for if 
we 2"o behind the cell we have structures which are not 
capable of leading an independent individual life. 

But a cell in the case of metazoa, or the nucleus and 
other structures in the case of protozoa, and unicellular 
plants are things which, whilst they participate in, and con- 
tribute to life, and to that extent may be considered as living, 
are not in themselves capable of independent individual 
existence. They may be called metabionts. 

The terminology suggested may not be perfect, but by 
the use of it or of something equivalent we may shake our- 
selves free of the false ideas which have clustered about 
individual life units, and start with a new hope on an inquiry 
into the nature and growth of bionts. 

An essential part of our conception of a biont is the 
union of two substances, cytoplasm and nuclein. It does 
not matter, for present purposes, that we know nothing 
exact about these two substances, and still less of the 
manner in which they operate together to produce the 
phenomena of life. It suffices that we know that there are 
bionts whose structure is so simple that we can affirm no- 
thing more of them than that they consist of cytoplasm and 
nuclein, e.g., Bacteria, Yeast, Oscillaria, etc. 


Within the limits of the protozoa we study many kinds 
of bionts which, whilst retaining great simplicity of structure, 
have advanced far beyond the stage represented by these 
simple forms. 

The most important as well as the most striking 
structural advance is the formation of a nucleus. The 
nuclein which was, in the simplest bionts, distributed 
through the protoplasm, is aggregated to form a compact 
body, which from its structure and behaviour may be re- 
garded as a metabiont, as also may the part from which it 
was segregated, the cytoplasm. The steps which lead up 
to the segregation of the nucleus are obscure, but there are 
very good grounds for saying that the nucleus, when formed, 
is connected, in some manner unknown to us, with the 
transmission of the so-called historic qualities of the biont. 
In any case it plays a leading part in reproduction, and the 
steps from the condition of diffused nuclein to centralised 
nuclein are suggested by the infusorian Holosticha scutelhim, 
which ordinarily has no definite nucleus, but contains 
numerous chromatin particles scattered throughout its sub- 
stance. Previous to reproduction by division the scattered 
particles are drawn together and unite to form a centralised 
nucleus, which divides in a normal manner and breaks up 
again into particles in the offspring. 1 

Besides the nucleus many other structural advances are 
to be noted in protozoa and in unicellular plants ; some 
must be regarded as metabionts, e.g., chlorophyll corpuscles 
and chromatophores of various kinds, many kinds of granules, 
etc. Other structures cannot be regarded as belonging to 
the same category, e.g., cilia, contractile fibres, etc. We 
may for the present purpose leave both cases out of con- 
sideration, for it is the nucleus and the part it plays as an 
essential constituent of the biont which most concerns us. 

We have as yet very obscure notions about the co-opera- 
tion of nucleus and cytoplasm in the production of vital 
phenomena. But, putting aside the views of those who 
postulate the existence of minute vital units, and speak of 

1 Aug. Gruber, " Ueber vielkernige Protozoa," Biol. Centralblatt, iv., p. 
170. See also the same author, Zeit.fiir Wiss. Zool., xli., p. 186. 



an emanation of specialised biophors from the nucleus into 
the cytoplasm, there is a general agreement that the co- 
operation is of the nature of a complex exchange of 
chemical material. If this be the case, the rate of exchange 
must be the measure of vital activity, and it is clear that 
the rate of exchange will be greatest in immediate proximity 
to the nucleus and will become increasingly less the greater 
the distance from the nucleus. At a certain distance, which 
might be called the limit of nuclear influence, the rate of 
exchange will be reduced to zero. We see that in the 
protozoa the forms which have a single nucleus are small, 
and we may say, in consequence of the foregoing- considera- 
tions, that their size is determined by the limits of nuclear 
influence. But many protozoa are multinuclear, and I 
believe that there is no exception to the rule that protozoa 
of relatively large size are also multinuclear. Such is 
obviously the case in such forms as Radiolaria, Actino- 
sphserium, Pelomyxa, the Myxomycetes and others. From 
a consideration of all the facts of the case we may legiti- 
mately infer that in any given biont growth beyond 
certain limits is incompatible with a uninuclear condition, 
and that further growth involves multiplication of the 
nucleus, which may have as consequences: (i) discon- 
tinuous growth, which in its simplest form is reproduction 
by binary fission : (2) continuous growth, in which the 
nucleus is multiplied so that all parts of the enlarged cyto- 
plasm may receive an equal share of nuclear influence. 
There are numerous cases in which, as I pointed out before, 
the two conditions are combined. There is a ccenocytial r 
stage of considerable duration, followed by reproduction 
(or discontinuous growth). 

The next phase is the formation of a biont of consider- 
able size, in which very numerous nuclei are arranged in 
definite manner in a continuous mass of protoplasm. Such 
a condition is represented by the Cceloblastas, and also in the 

1 When in my earlier essay I coined the word hypopolycytial I was 
not aware that Professor Vines had applied the term ccenocytial to the 
Cceloblastse. His term has the priority and is more euphonious, so I adopt 
it instead of my own. 


growing tissues of many animals and plants, as for instance 
in the embryoes of many Arthropods, in the endosperm of 
Phanerogams, etc. The condition may be permanent, as 
in the case of the Cceloblastae, or non-permanent, as in the 
other cases. But in both instances there is a difference 
from the ccenocytial condition observed in Protozoa, namely, 
that the multiplication of the nucleus does not lead to re- 
production in the form of the splitting up of the biont into 
as many new bionts as there are nuclei. 

In a ccenocytial biont of appreciable size the relations 
of the various parts to external conditions will tend to be- 
come different, and differences of chemical constitution will 
be set up in the different regions exposed to different con- 
ditions. We can see that this is the case in Botrydium, in 
which root and shoot are plainly marked off from one 
another, and better still in Caulerpa, in Codium, and in 
many of the Moulds. Differences in chemical constitution 
thus induced will mean difference in exchange between 
nucleus and cytoplasm, and we may infer that, in accordance 
with these differences, the cytoplasm within the limit of 
influence of any one nucleus will in time assume a con- 
stitution so different from that of the adjacent cytoplasm as 
to become sharply marked off from it. It will then acquire 
its own surface tension — the first step towards a cell wall— 
and will be a separate corpuscle containing a nucleus, in 
fact a cell. Such a cell however has not come into being as 
an individual unit joined to its like, either phylogenetically 
or ontogenetically, but it has from the first formed a part 
of an organic whole, of which it is nothing more than a 
specialised component part. 

One looks naturally for evidence of this mode of forma- 
tion of cellular structure in developing Metazoa. The best 
evidence is to be found, I think, in the segmentation and 
formation of the layers in many Ccelenterata. In some 
Ccelenterata — for example, in Renilla — the nucleus divides 
without accompanying division of the cytoplasm until eight 
or sixteen nuclei are present, and then the cytoplasm 
divides and eight or sixteen cells are formed. But of more 
importance than this is the formation of the layers. From 


the considerations stated above we should expect that the 
changes in chemical composition of the cytoplasm and the 
correlated changes in the nucleus, in other words the dif- 
ferentiation, would first become manifest in the peripheral 
parts of the growing ccenocyte, and that we should have a 
stage in which there was a cellular external layer and a 
ccenocytial internal mass. We find that in fact in the 
embryoes of many Ccelenterates the outer layer is divided 
up early into sharply defined cells at an early period, whilst 
the central cells retain the character of a ccenocytium ; at 
most the cell outlines of the internal mass are confused and 

We see also that in the growing tissues of the embryoes 
of higher animals the embryonic tissue is not cellular but is 
a ccenocytium, for example, the mesoblast of Avian and 
Selachian embryoes and of the Rabbit. It is only at a later 
stage when different relations to other parts of the body 
have been acquired and new exchanges of material are 
forced upon the growing mass, that the continuous mass of 
cytoplasm is split up into corpuscles, each of which, in my 
view, corresponds to the limit of influence of a nucleus. 

On the other hand we have the undoubted fact that in 
many organisms there is no ccenocytial phase in develop- 
ment, but the cytoplasm surrounding the nuclei as they are 
successively formed is immediately marked off into definite 
corpuscles, so that the whole process of development 
suggests the formation of an aggregate of bionts derived by 
division from a single parental biont. An explanation of 
this fact presents many difficulties, and I have not now 
the space to discuss these difficulties and to show that, 
obscure as the subject still is, there is ground for supposing 
that the difficulties are chiefly due to the prepossession 
which exists in most minds in favour of the independent 
life unit theory. I hinted in my previous paper (loc. ciL, p. 
171) that the discrete condition of the blastomeres of so 
many embryoes may be connected with the fact that they 
are, from the very outset, specialised. This means that as 
the nucleus is in some way associated with the transmission 
of historic qualities, these qualities may be located in special 


parts of the nucleus, and on division, some of the daughter 
nuclei may possess one set, others may possess another set 
of " qualities ". By " qualities " I conceive that we mean 
different chemical constitutions, and it would follow that the 
daughter nuclei, being of diverse chemical constitutions, 
would react in diverse manners on the adjacent protoplasm 
and would each cause the delimitation of a territory of 
cytoplasm within the limits of its own sphere of influence ; 
in other words, cell bodies would be formed round nuclei of 
different chemical constitutions. 

There is, however, yet another consideration to be taken 
into account. As Hertwig has shown, the cytoplasm in 
many ova is not homogeneous but is obviously separable 
into tracts of unquestionably different chemical constitution. 
This is conspicuously evident in the ova of Amphibia. As 
the nucleus divides, its products come into relation with 
different kinds of cytoplasm and the exchanges between 
nucleus and cytoplasm will be different in different places 
within the limits of the egg. Arguing on the same prin- 
ciples as before, we may attribute the successive formation 
of discrete blastomeres to this factor as much as to the 
separation in the course of division of different qualities 
contained in the egg nucleus, and according as one leans 
towards an epigenetic or an evolutionary theory of develop- 
ment so will one be disposed to lay more stress on the 
one factor or the other. There is this much to be said, 
that the most remarkable cell-lineages (which are only 
traceable when the blastomeres are discrete) have been 
observed in ova which contain a considerable proportion of 
yolk, which is not evenly distributed throughout the egg, 
and it is suggestive that segmentation in all cases leads to 
the segregation of corpuscles richer in yolk from corpuscles 
poorer in yolk — in fact to the segregation of materials of 
diverse chemical constitution. 

Tempting as it is to pursue this subject further, I must 
not attempt to do it now. But as I have claimed that the 
views which I have tentatively put forward are agreeable 
to the facts which we are in possession of, I may well give 
a short summary of the facts which I have relied upon. 


( i ) The co-existence of two substances at least, nuclein 
and cytoplasm, is requisite for life. (This is an inference, 
strictly speaking, and not a fact ; but I think that it may be 
considered a legitimate inference from what we know of the 
structure of the lowest bionts, and from the experiments of 
Nussbaum, Gruber, Verworn and others.) 

(2) The existence of bionts, such as Bacteria, in which 
we are unable to distinguish more than these two sub- 
stances. (This is a fact, which lends material support to 
the above inference.) 

(3) The existence of bionts in which nuclein and cyto- 
plasm are not indefinitely intermingled, but the former is 
segregated in the form of particles scattered through the 
protoplasm, e.g., Trachelocerca phcenicopterus and Chcenia 
teres. (We gather from this fact that the two chemical 
substances tend to become separated from one another.) 

(4) The temporary aggregation of nuclein particles to 
form a centralised nucleus for the purpose of the repro- 
ductive act, e.g., Holosticha scutellwm. (We infer from this 
that there is some connection, at present hidden from us, 
between the nucleus and the reproductive act.) 

(5) The existence of many bionts in which the nuclein 
is concentrated to form a nucleus. (We infer that this 
is a grade of permanent differentiation arising out of 
the previous temporary grade.) 

(6) The existence of many nuclei in all bionts which, 
whilst still undivided as regards their cytoplasm, attain to a 
certain size. (From this we infer that the " limit of nuclear 
influence " cannot extend through a large mass of cyto- 

(7) The origin of "cellular" tissues from a ccenocytial 
mass, e.g., the endosperm of Phanerogams; the neural 
crest of certain Vertebrate embryoes ; the embryoes of 
Arthropods; the mesoblast of many Vertebrates, etc. (From 
this we infer that the cells composing many tissues of higher 
animals are not to be regarded as bionts, but are secondarily 
derived during the growth and extension of the parts of a 
single biont.) 

This re'sume suffices I think to show that this at least 


may be claimed for the views which I have put forward. 
They are founded strictly on the facts, and they do not 
depend on the assumption of any kind of hypothetical units 
of which the nature and even the very existence is entirely 
beyond our ken. 

Since I have not been able to develop my views, I 
cannot but expect that they will be subject to considerable 
modification and even to entire overthrow. They form at 
least an attempt to classify and colligate the various pheno- 
mena which seem to be germane to the subject, and I have 
collected and compared a much larger body of facts than I 
am here able to refer to, without finding any which are 
contradictory to my ideas. That my ideas are somewhat 
indistinct need not, at present, be urged as an objection, for 
indistinctness is not necessarily a sign of falsity. The cell- 
republic theory was not wanting in distinctness, but 
it was inappropriate to the facts. I only claim that my 
ideas are appropriate, and I shall hope to give them more 
distinctness on a future occasion. 

In the meantime I leave out of consideration a large 
question, concerning which I think it scarcely possible to 
give a satisfactory account, in this standing in opposition to 
Mr. Sedgwick, who thinks that which I have attempted to 
be impossible, but offers a solution of that which I think 
scarcely possible. 

The question is, how are we to account for that pheno- 
menon which I have described as a progress from the state 
of an independent corpuscle, through a state of many coherent 
or continuous or conjunct interdependent corpuscles, back 
again to the state of a single independent corpuscle ? 

Mr. Sedgwick's solution is this : that the unicellular 
form is assumed by metazoa in order that conjugation may 
be possible. The single independent corpuscle which re- 
curs in the cycle is the sexual cell, and the essential feature 
of sexual reproduction is the conjugation of reproductive 
cells. The unicellular phase is only assumed in sexual, not 
in asexual reproduction, and this is to be explained by the 
consideration that conjugation is as necessary in metazoan 
life as in protozoan life, but that conjugation between the 


ordinary forms of metazoa is impossible for mechanical 
reasons, and therefore special individuals of a form simple 
enough to admit of conjugation are produced. These 
special individuals are the ovum and spermatozoon. 

The explanation is extremely ingenious and there is 
nothing unreasonable in it, but one cannot say that it is 
altogether acceptable at first sight. It would have been 
more satisfying if Mr. Sedgwick had marshalled some of 
the facts relative to the sexual reproduction of some of the 
lowest multicellular organisms and had shown their rela- 
tion to his suggestion. A difficulty which at once occurs 
to me is that in many plants asexual reproduction is 
effected through the agency of a single cell. In fact, 
before one can accept any solution of the question one 
requires a very extensive and careful survey of all the facts 
known about the reproduction of the lower plants. They 
afford examples of every conceivable grade of the reproduc- 
tive processes, and, once one begins to look into the subject, 
hints as to the parting of the ways of sexual and asexual 
reproduction occur to one at every step. The pity is that 
the mere zoologist, who does not find such a fruitful field in 
his own territory, is obliged to disinter the facts from the 
load which the peculiarities of botanical terminology have 
heaped upon them. 

It is quite possible, however, that such a survey would 
afford strong support to Mr. Sedgwick's opinions, and if it 
should do so they would in no way be inconsistent with 
the ideas which I have put forward, but would rather sup- 
port them. 

A word in conclusion for those who will reproach me 
for having attempted to frame a chemico-physical theory 
of organic growth, and for having used such phrases as 
" complex chemical constitution," " exchange of chemical 
material," etc., without assigning any distinct meaning 
to them. I admit that our knowledge on the subject 
is rather inadequate, and that I have used obscure phrases 
to express relations which are in themselves obscure. If 
one attempts to lift the veil of obscurity one must inevitably 


call hypothesis to aid, and it has been my object to avoid 
the use of hypothesis where I could do without it. It is, 
however, legitimate to frame an argument which, while it 
agrees with the lessons of experience, is ultimately based 
upon hypothetical considerations, provided always that those 
considerations are consistent with the accepted teaching of 
the sciences whose aid is invoked. 

Any attempt whatever to find an explanation of vital 
phenomena ends in an appeal to chemistry and physics. 
Knowing as we do that the elements of which organic 
bodies are composed are not different from those which 
occur in the inorganic world, we cannot refuse to acknow- 
ledge that vital processes are in the end chemico-physical 
processes, and this much is conceded by every author of a 
theory of vital units. The difficulty which they have to 
face is the same as that which I have to face, and is not one 
whit the less because it is compressed into the limits of 
a biophor, whereas I would allow it the limits of a cell. 
Can we frame any distinct ideas of these chemico-physical 
processes ? Not very distinct ideas, perhaps, yet we can 
supplement the lack of positive evidence by analogies and 
illustrations involving the same ideas as those which are 
current in the physical world. 

It was Professor W. K. Clifford, I think, who first drew 
a graphic picture of the molecular forces which are at work 
in any chemical compound, by describing the atoms as 
linked to one another and dancing a sort of merry-go-round 
within circumscribed limits. We may carry on the illustra- 
tion, which, fanciful though it may seem, is supported by 
physical and mathematical considerations. A biont is a 
great organised war dance, performed by a whole army 
corps. The individuals composing each company are the 
atoms, they are linked to one another by companies and 
each company dances its own figure. Every company is a 
molecule, and every company dance is but a part of a larger 
dance, in which the companies act in relation to one another 
as the individuals act in the company dance. The larger 
dances are regimental dances and every regiment is a 
micella. The regimental dances are but parts of still larger 


brigade dances, and the brigade dances are but part of the 
great dance of the whole army corps, which, taken as a 
whole, is the biont. The illustration is not quite exact, for 
each company must not be considered as consisting of like 
individuals, but of many individuals of all arms, some like 
and some unlike, linked in such various ways that no two 
companies are the same, partly because of the proportions of 
different kinds of individuals composing them, partly because 
of the way in which those individuals are linked together. 
Nor must we imagine that individuals are permanently 
attached to companies, nor yet companies to regiments, but 
that in the course of the dance individuals are passed from 
company to company, and companies from regiment to 
regiment, each conforming temporarily to the particular 
figure of that part of the dance to which he or it for the 
time belongs. Further than this the individuals engaged 
in the whole dance are never lone the same : there are 
bystanders who for a time do not participate in the dance 
but are caught up one by one, whirled through the figures, 
passed from company to company, from regiment to regi- 
ment and brigade to brigade, and are eventually passed out 
of the dance again, after having participated in some or all 
of the figures as the case may be. Every individual in the 
dance is at some time passed out of the dance, becomes 
a bystander, and may again be caught up and whirled along 
in the dance once more. 

The illustration is farfciful, if you please, but it is of the 
same kind as illustrations used to depict the play of mole- 
cular forces in the inorganic world. It serves a purpose in 
that it gives the imagination something to work upon, and 
it enables one to conceive of the immense complexity which 
is possible in a chemico-physical process. The army dance 
which I describe is capable of any number of combinations, 
a number amply sufficient to satisfy the needs of those who 
insist so strongly on the marvellous complexity of life. 
Let anybody imagine an army to be composed of four 
brigades, each brigade of four regiments, each regiment of 
ten companies, and each company to contain 100 indi- 
viduals of the eight kinds, carbon, oxygen, hydrogen, 


nitrogen, sulphur, phosphorus, potassium and iron, in 
varying proportions, and let him work out the possible 
combinations. I think he will be satisfied with the com- 

What then of heredity and of the capacity which I have 
mentioned for acquiring historic qualities ? 

Believing as I do that the vital processes must in the 
end be attributed to a particular mode of molecular motion, 
I believe that it is the form of movement which is trans- 
mitted. Returning to my illustration I would say that it is 
the figure of the whole dance which makes up the species, 
and that it is the figure — the mode of motion — which is 
inherited, clearly not the individuals engaged in the dance, 
except in a very small degree, for they are constantly 
coming into the dance anew and as constantly being passed 
out of it. Under certain circumstances there may be an 
excess of one or more kinds of new individuals pressing into 
one part of the dance which will affect the figure of the 
company dance which they crowd into, and this will affect 
regimental figures and ultimately, in decreasing degrees, 
the whole army figure. In this way we may picture to 
ourselves the action of external influences in bringring about 
variation. But I have given rein enough to my imagina- 
tion. The picture was introduced partly to show that 
beneath my obscure phrases there was some distinctness of 
ideas, partly to emphasise the immense complexity of 
Nature and to show that even atoms and molecules may be 
conceived to be so combined together that, in Goethe's 
words, " sie bewirken so eine unendliche Production aut 
alle Weise und nach alien Seiten ". 

Gilbert C. Bourne. 


IT is well known that in the construction of many of the 
theories of heredity the doctrine of the transmission of 
acquired characters has obtained considerable prominence. 
The hypothesis of Lamarck rendered it necessary to assume 
that structural characters which had arisen from the use or 
disuse of organs, became an integral part of the individual 
and reappeared in the descendants, and although the appli- 
cation of this idea became greatly restricted when the 
principle of natural selection was established, it is only 
within the last few years that the transmission of acquired 
characters has been considered as unproven, and the in- 
stances put forward in support of this view to be capable 
of a different explanation. It may be admitted that mutila- 
tions and permanent injuries can be included among acquired 
characters, and the structural and functional modifications of 
the individual which occur in disease may persist, and 
therefore also be considered as definite morphological or 
physiological changes. Mutilations apparently do not pass 
from parent to offspring, and this has been especially pointed 
out by Weismann and his followers, since, if heredity is 
capable of explanation on the hypothesis of the continuity 
of germ-plasm contained in definite reproductive cells, any 
change in the structure or modes of activity of the essential 
body or somatic cells would not be transmitted. An iden- 
tical line of argument also negatives the belief that diseases 
can be inherited, and this view was maintained by Weismann 
in his well-known criticism on the transmission of experi- 
mental epilepsy ; the symptoms in this hereditary disease he 
considered might be due to some unknown microbe which 
found its nutritive medium in the nervous tissues and 
was transmitted in the reproductive cells. The question 
whether micro-organisms can actually pass from parent to off- 
spring is one which has been carefully investigated, whereas 


the proof that actual morphological changes, such as modi- 
fications of histological or molecular structure, can be trans- 
mitted has not yet been given. It is conceivable that 
predispositions may be inherited, and these must result 
from alterations in the germ-plasm, or a direct infection of 
the germ or embryo might cause the transference of a dis- 
ease from one generation to another, a phenomenon which 
simply depends upon a particular mode of conveyance of a 
parasite. 1 

In many diseases, and particularly those which are directly 
caused by micro-organisms, it is a matter of interest to note 
the wide differences which exist between the conveyance 
of hereditary characters, and of a specific disease. Armauer 
Hansen (1) has made this perfectly clear in considering the 
etiology of leprosy. He has pointed out that true heredi- 
tary characters are usually limited to one sex, frequently 
appear at a particular age, and the phenomenon of atavism 
is not rare ; but in the conveyance of such a disease as 
tuberculosis or leprosy, none of these conditions are ful- 
filled. It is a logical deduction from the consideration of 
these differences that every specific disease which is trans- 
mitted cannot be regarded as hereditary, but as an instance 
of the direct bacterial infection of the germ-cells or embryo. 
Most writers on cancer and malignant growths have dis- 
cussed the hereditary transmission of this disease, and if it 
is allowed that a disposition to cancer may be derived by 
inheritance, then this condition would depend upon some 
peculiarity inherent in the nucleus of the germ-cells ; but 
if, on the other hand, malignant disease is caused by a 
parasite belonging, as some investigators have sought to 
prove, to the group of protozoa or protophyta, then the 
transmission of the actual disease will depend upon the 
passage of a micro-organism which invades the germ or its 

1 " Pour les maladies, vraiement constitutionnelles, c'est la substance 
hereditaire elle-meme qui est viceuse; pour les maladies infectieuses, levice 
n'est pas dans la substance elle-meme, mais a cote d'elle, et les produits 
sexuels servent seulement de vehicule a un parasite capable d'engendrer 
plus tard une maladie generate. " Y. Delage, La Structure du Protoplasma 
et les Theories sur F Heredite. Paris, 1895. 


product, and the whole phenomenon ceases to be one of 
heredity, for the hereditary transmission of micro-organisms 
is simply a particular instance of bacterial infection. The 
inheritance of actual specific disease must therefore always 
be considered as a problem absolutely distinct from that 
of heredity and incapable of explanation by any hypothesis 
of heredity. 

Micro-organisms which reach an individual either by 
inheritance or other modes of conveyance may undoubtedly 
exhibit a period of latent life extending over many years ; 
but when this condition is succeeded by an active life, to 
establish the proof of an hereditary transmission is ex- 
ceedingly difficult or even impossible (11). The early 
researches into problems of this nature were necessarily 
made with the help of statistical and clinical methods ; but 
it is the application of experimental methods, which could 
only be pursued with success as the study of bacteriology 
developed, that has finally succeeded in removing the subject 
of the hereditary transmission of specific diseases from the 
hazy region of speculation. The attitude assumed by 
Baumgarten and his followers on this question is well 
known. In the case of tuberculosis it is maintained 
that individuals are rarely infected with tubercle bacilli 
after birth, but that the disease in the majority of cases is 
due to a parasitic infection of the egg-cell or embryo. It 
is even urged that the bacilli may remain latent in one 
individual, and only enter upon a phase of activity in the 
offspring, a view which, if correct, would accord with the 
opinion of many clinical observers. While destroying the 
opinion so commonly held that an "inherited tubercular 
predisposition " exists, Baumgarten's theory of hereditary 
parasitism makes a still greater demand on the imagina- 
tion ; but that the views of this distinguished pathologist 
have acted as a stimulus to renewed experimental work on 
the transmission of micro-organisms is beyond doubt. 
Recent papers by O. Lubarsch (2) of Rostock and J. 
Csokor (3) of Vienna give an admirable exposition of the 
present position of our knowledge on this subject of the 
transference of bacteria from parent to offspring in man and 


the lower animals, and the evidence that bacteria may in 
this manner gain access to the organism is incontestable. 

In inherited specific diseases it is possible to distinguish 
two forms of infection : first, by a direct bacterial invasion of 
the essential reproductive cells ; secondly, the egg-cell or 
the embryo may receive micro-organisms from the female, 
in which case the blood stream is the channel for conveyance, 
and the whole phenomenon is then one of metastasis com- 
parable in every respect to what obtains when bacteria 
multiply at a definite area of the body, and thence become 
distributed by the blood and lymph in distant parts of the 
organism. Bacterial infection may therefore be either 
germinative or placental, and in mammals the latter 
form of transmission is not infrequently observed. The 
specific bacteria of anthrax, typhoid fever (6), pneumonia 
and tuberculosis (7) have been isolated from the human foetus, 
cultivated, and successfully inoculated upon animals, so 
that the chain of evidence is complete. The pyogenic 
cocci such as streptococcus pyogenes (24) and staphylococcus 
pyogenes aureus have also been demonstrated in foetal 
tissues by Fraenkel and Kiderlen, and Auche has shown 
that in small-pox the placenta may be penetrated by these 
micro-organisms. In the lower animals not only may the 
bacteria already mentioned be transmitted, but also those of 
cholera, glanders and chicken cholera. 

In many animals the egg-cell is the largest unit of the 
organism, and would be capable of containing numberless 
bacteria ; that such an infection does occur was first estab- 
lished by the classical observations of Pasteur (4), which 
have been confirmed by all subsequent investigators. In 
pebrine, a disease of silk-worms, definite sporocyst forms 
(microsporidia or Cornalia's corpuscles) are transmitted from 
the imago in the egg-cell, and the larva is directly infected 
in this manner. Blochmann (5) has also described a similar 
mode of conveyance of bacteria in the ova of Blatta 
orientalis. In a single instance a tubercle bacillus has 
been seen in the mammalian ovum. The sperm-mother- 
cells may also be invaded by micro-organisms, but this is 
rare, and no example of an infected male reproductive cell 


exists. That this condition will ever be demonstrated is 
improbable, since bacteria contrast with parasitic protozoa 
in infecting the cell and sparing the cell-nucleus, and the 
essential agent in the process of fertilisation is the nucleus 
or head of the sperm-cell. 

Various observers have attempted a solution of this 
question of germinative infection by the employment of 
two different methods. The first of these is that pursued 
by Mafifucci, who directly infected the fertilised eggs of the 
fowl, and in the second not only were the genital glands 
and the products of these examined for micro-organisms, 
but pieces of them were taken from animals suffering with 
specific diseases and used as material for inoculation. 

Even if it is assumed that an ovum actually is a site in 
which bacilli such as those of tuberculosis exist, it may be 
objected either that the microbe is dead, or that such a cell 
is incapable of development. This is the attitude taken by 
Virchow, who absolutely denies the existence of congenital 
tuberculosis. Maffucci's experiments, however, contra- 
dict this opinion, for this observer has shown that the 
bacilli of avian tuberculosis develop in an infected embryo, 
and the chicken succumbs to tuberculosis in twenty days 
to four and a half months after hatching. If, however, 
instead of infecting the embryo, bacteria such as those of 
chicken cholera, or anthrax, or Friedlander's pneumococcus 
are introduced in the extra-embryonic area, then these 
organisms may actually enter the embryo through the 
allantois but do not increase in number provided the 
embryo remains alive. The pathogenic micro-organisms 
may therefore be destroyed or attenuated by actively pro- 
liferating embryonic tissue cells, or they may become capable 
of development at a later period of life, in other words, 
remain latent. Although these experiments were devised 
to establish the view that a genuine germinative infection 
may occur, they obviously do nothing of the kind, and it is 
to the researches of Gartner that we owe an absolute 
demonstration that ova may contain pathogenic germs. 
Gartner among: other animals inoculated canaries with mam- 
malian tubercle bacilli. After a few weeks he removed 


nine eggs, washed these in dilute corrosive sublimate, 
dried them in wool and introduced the contents of each eo-or 
into the peritoneal cavity of guinea-pigs. In two cases 
tuberculosis was set up, the animals dying one and a half 
months and two and a half months after infection. These 
experiments, which are absolutely free from objection, 
conclusively prove that the egg-cell may contain virulent 
bacteria, and it is easily conceivable that such eggs may 
develop and the transmission of the parasite take place by 
direct germinative infection, especially since Maffucci's 
work shows that such infected eggs are capable of develop- 

Jani, Westermayer, Spano, Walther, Gartner, and quite 
recently Jakh, have microscopically investigated the bac- 
terial contents of the reproductive glands, and also inoculated 
animals with fragments of these organs. With the exception 
of Gartner's researches these experiments have not added 
greatly to our knowledge of the hereditary transmission of 
bacteria. All the experiments of Westermayer were nega- 
tive. In fourteen cases of well-marked General tuberculosis 
no tubercle bacilli could be recognised, and inoculation 
experiments were failures. The experiments of Jakh (10) 
were more fortunate. Five inoculations with pieces of the 
male reproductive gland and its product, taken from in- 
dividuals dead of tuberculosis, gave three positive results. 
If the gland alone was used, the experiments were always 
negative, and of three inoculations with pieces of the egg- 
forming gland one was successful. It may be admitted that 
these experiments do not really throw much light on the 
subject of germinative infection, but Gartner's researches 
are of much greater value. He experimented upon mice, 
guinea-pigs, rabbits, and canaries, these birds being sus- 
ceptible to mammalian tubercle bacilli. Having inoculated 
these animals with bacillus tuberculosis, a careful examination 
was made of the offspring of such tubercular parents. This 
method might naturally be expected to give a conclusive 
answer to the question of hereditary infection, and the 
following information has been gained from these researches : 
1. The sperm rarely contains tubercle bacilli — five in 



thirty-two cases. Even if micro-organisms exist they are 
incapable of infecting the egg. In twenty-two (rabbits) 
and twenty-one cases (guinea-pigs) where the male repro- 
ductive gland was the seat of an acute tubercular process, 
the offspring were never infected. 2. Neither does the 
male infect the female by way of the sperm. 3. Infection 
takes place frequently from the female to the foetus, and in 
an overwhelming majority of cases by way of the placenta. 

A few considerations may make the importance of Gart- 
ner's work more evident. If bacilli exist, as they occasionally 
do, in the product of the male gland it is probable that this 
material, like other parts of the body, contains bacteria only 
a few days before death, for we know that quite an abnormal 
number of micro-organisms may invade the whole organism 
during the last days of life. Tubercle bacilli are immotile 
and therefore will not easily reach the oviduct or egg, a 
matter of some importance, since it has been shown that in 
most cases the ovum is fertilised either high up in the 
oviduct or even at the time of liberation from the Graafian 
follicle. Stroganoff (12) has also pointed out that the 
uterine area is sterile, and the secretion of this is bacteri- 
cidal, in which it resembles sputum (13) or the mucus of the 
nasal tract which is almost free from germs (14). Lastly, it is 
well known that a single male morphological unit is sufficient 
for fertilisation, and if we assume with Gartner that 100 
virulent tubercle bacilli are mixed with sperm-cells, the 
ratio of bacteria to these would be about 1 : 22,500,000 ; it 
is hardly conceivable on the doctrine of probabilities that a 
bacillus would gain access to the egg. It may therefore be 
considered, both on experimental and theoretical grounds, 
that a germinative infection of the ovum never occurs by 
the conveyance of micro-organisms in the male reproductive 

The difficulties which exist in proving that the in- 
heritance of a specific disease may occur through an in- 
fection of the ovum are fortunately not so great in those cases 
where the passage of micro-organisms takes place solely 
from the female to the fcetus by way of the placenta. It is 
established that specific micro-organisms can pass by this 


route. It is not even necessary to assume that there is any 
lesion whatever in the placenta or that the epithelium of 
the foetal villi is destroyed. An experiment by Zuntz 
shows clearly that particulate material will easily pass into 
the amniotic fluid from the maternal portion of the 
placenta, for if indigo-carmine is injected into the veins 
of the female the dye passes into the amnion leaving the 
foetus free, and in this very manner anthrax bacilli may 
pass, and from the amnion gain access to the mouth of 
the foetus, enter the gut and set up disease by a 
primary infection of the wall of the intestine (25). An intra- 
uterine infection, therefore, can occur without lesion of the 
placenta, though in the majority of cases this structure is 
primarily infected, and then the foetus, or else haemorrhages 
of the placenta permit the passage of micro-organisms. 
However the undoubted fact that micro-organisms can 


penetrate the healthy skin by way of the hair follicles — and 
the same is possibly true for the epithelium of the urinary 
tract — must not be forgotten in considering the passage of 
bacteria across the placenta. This structure may be nor- 
mal and even then allow the transit of bacteria. Birch- 
Hirschfeld (15) from researches on the structure of the 
human placenta as well as that of mice, rabbits and goats 
considers that the bacilli of anthrax at any rate can 
traverse the uninjured chorionic epithelium. Moreover in 
the human placenta and in rabbits numerous processes of 
the chorion traverse the placental sinuses, and these pro- 
cesses are normally destitute of epithelium. It was noticed 
by Max Wolff (16) that anthrax bacilli easily pass if the 
placenta was crushed or torn, and micro-organisms which 
exert a necrotic influence on tissues, such as the pyogenic 
cocci, appear first to destroy the epithelium of the 
chorionic villi, and then pass through into the foetal 
blood. In this fluid micro-organisms reach the liver, and 
it is this organ which, as a rule, is primarily affected, and 
then the glands in the lymphatics leading from the organ 
become implicated. The location, therefore, of tubercles 
in foetal tuberculosis is characteristic, and all observers 
insist upon this feature in determining whether tubercular 


deposits are of intra- or extra-uterine origin in early cases 
of the disease. As a matter of interest it may be 
mentioned that quite recently Bar and Renon have de- 
monstrated tubercle bacilli in the blood of the umbilical 
vein (7). The method used by these observers, that of 
inoculating guinea-pigs with the suspected blood, and in 
this manner establishing tuberculosis, is not so convincing 
as the actual demonstration of bacteria in fcetal tissues. 
Wassermann (17) in a recent paper especially insists on 
this point, and discards all evidence of inherited disease 
which rests simply upon inoculation experiments. He 
describes a case of early tuberculosis which ended fatally 
when the child was ten weeks old, where the disease was 
acquired, not from the parents who were healthy, but by 
direct infection from a tubercular relation, and believes that 
such cases as these are not infrequently cited as instances 
of congenital disease. In his opinion hereditary trans- 
mission of bacteria does occur, but it is exceedingly rare 
in comparison with the frequency of extra-uterine infection. 
Bernheim (18) considers that the offspring rarely, if ever, 
become tubercular if separated from tubercular parents, 
with the exception of those cases where the placenta is 
infected. The case reported by Ivan Honl (19) of a child 
fifteen days old that on autopsy showed tubercular nodules 
in the liver, spleen, and lungs, and numerous bacilli, 
must be classed as a definite case of transmission which with 
many others lends no support to Eberth's statement that 
individuals do not inherit tuberculosis but acquire it (23). 

A recent case of congenital typhoid fever is related by 
Freund and Levy (20), and instances of transmitted hemor- 
rhagic infection have been recorded by Neumann (21) and 
by Dungern (22). The numerous examples which the 
journals of veterinary science contain, especially the work 
of Bang, Kockel, and Lungwitz, also afford conclusive evi- 
dence of the transmission of pathogenic micro-organisms, 
though there is a consensus of opinion that the placental 
is far more frequent than the germinative infection. The 
share borne by the male in this transmission may be dis- 
regarded, as no bacteriological evidence exists to support 


this view. Finally, the frequency of hereditary transmission 
of pathogenic germs is exceedingly small compared to other 
modes of infection. 


(1) HANSEN and LOOFT. Leprosy in its Clinical and Pathological 

Aspects, 1895. 

(2) Lubarsch. Ergebnisse der allgemeinen Atiologie der Mens- 

chen-und Tierkrankheiten, by Lubarsch and Ostertag, p. 427, 
1896. References to the transmission of infectious diseases 
to descendants will be found in this and the following paper. 
Some additional and later references are given in the course 
of this article. 

(3) Csokor. Ibid., p. 456. 

(4) Pasteur. Etudes sur les maladies des vers a soie, t. i.,'p. 70, 


(5) BLOCHMANN. Quoted by L. Pfeiffer in Die Protozoen als 

Krankheitserreger ; 1 89 1 . 

(6) Janiscewski. Munch, med. Wochenschrift, 1893. 

(7) Bar and Renon. Comptes Rendus, No. 23, 1895. 
Londe. Comptes Rendus, No. 25, 1895. 

Nocard. Un nouveau cas de tuberculose congenitale. Rev. 
de Tuberculose, No. 3, 1896. 

(8) MAFFUCCI. Centralbl. f. Bakt. u. Parasitenkunde, Bd. v., No. 

7 ; and Centralbl. f. allg. Pathologie, No. 1, 1894. 

(9) Gartner. Zeitschrift f. Hygiene, Bd. xiii. 

(10) Jakh. Virchow's Archiv, Bd. cxlii., 1895. 

(11) Washbourne and others in the discussion on latent micro- 

organisms at the Medico-Chirurgical Society, London. 
Lancet, November, 1895. 

(12) Stroganoff. Centralbl./. Gyndkologie, No. 38, 1895. 

(13) Sanarelli. Centralbl. f Bakt., Bd. x., 1892. 

(14) Hewlett. Lancet, June, 1895. 

(15) Birch-Hirschfeld. Zieglers Beitr. z. path. Anat.u. allg. 

Path., Bd. ix., 1891. 

(16) Wolff. M. Intemat. Beitr. z. wissensch. Med. Festschr. f. 

R. Virchow, Bd. iii., 1891. 

(17) Wassermann. Zeitschrift f. Hygiene, Bd. xvii., 1894. 
KOSSEL, H. Zeitschrift f. Hygiene, Bd. xxi., 1895. 

(18) Bernheim. Erblichkeit und Ansteckung der Tuberculose. 

Mitteilungen aus dem xi. internat. med. Kongresse in Rom, 
1894. Reference in Centralbl. f Bakt., No. 17, 1894. 


(19) Honl. Uber kongenitale Tuberkulose. Reference in Centralbl. 

f. Bakt., Bd. xviii., 1895. 

(20) FREUNDand Levy. Berliner klin. Wochenschrift,No. 25, 1895. 

(21) Neumann. Archiv f. Kinderheilkunde, Bd. xiii. 

(22) DUNGERN. Centralbl. f. Bakt., Bd. xiv., 1893. 

(23) Eberth. Die Tuberculose, ihre Verbreitung und Verhiitung, 

1 891. 

(24) RlCKER. Centralbl. f. allg. Path. v. path. Anat., Jan., 1895. 

(25) KOCKEL and LUNGWITZ. Beitr. z. path. Anat. v. allgem. 

Path., Bd. xxi. 

George A. Buckmaster. 

Science progress* 

No. 29. July, 1896. Vol. V. 


THE purpose of these notes is to summarise the results 
of recent research among the prehistoric peoples and 
civilisation of the Eastern Mediterranean ; especially in so 
far as these prepare the environment for the first great 
civilisation of Europe, namely, that of Greece, and fill the 
chronological gap, and explain such communication as 
existed, between this and the equally " historic " but far 
earlier civilisations of the Euphrates and Nile Valleys. 

A strictly " Historic " Age on the shores of the ^gean 
Sea, or in fact in the Eastern Mediterranean at all, cannot 
be said to begin before the seventh or at earliest the end of 
the eighth century B.C. ; and everything before this point 
would certainly have been classed as " Prehistoric," but for 
the fact that, until quite lately, the preceding centuries have 
been interpreted wholly in the light of a voluminous Greek 
tradition, which is still accepted in many quarters as 
fundamentally historical ; though now with wide reserva- 
tions everywhere. Consequently prehistoric archaeology 
and ethnology have here come into existence as accessory 
and supplementary studies, and the data of the literary 
tradition have been used, as was inevitable, as a working 
hypothesis ; which, it is only fair to say, has served its purpose 
fully as well as there was every reason to expect. Con- 
sequently again, any account of the more recent and more 



strictly anthropological work in this field must stand, if it is 
to be intelligible, in close relation with the data and 
assumptions, which have so mainly determined its course. 


i. The data upon which Greeks of the sixth and early 
fifth centuries relied for the reconstruction of their own 
history consisted wholly of traditional anecdotes, appended 
to traditional genealogies, or grouped, in more or less organic 
connection, round equally traditional events, such as an 
invasion of the Troad, or an exploration of the Euxine, or 
the adventures of a typical navigator like Odysseus. Many 
of the lays in which these anecdotes were preserved can be 
traced with some probability to their places of origin, which 
range from Cyprus to the islands off the west coast of 
Greece, and from Thessaly and the Troad to Crete. All 
profess to represent the civilisation of the yEgean area at a 
period removed by several centuries from the point at 
which the Hellenic world emerges into history ; and the 
traditional chronology of historical Hellas went up to an 
era which is slightly later, but approximately contemporary 
with the latest episodes of the Epic poems. Now though the 
lays which display the greater literary skill and the maturer 
idiom give a less vivid and more conventional picture ; and 
though occasional allusions occur to customs and beliefs 
which are characteristic of Hellenic culture, those others 
which Greek tradition reckons primary, namely, the Iliad 
and the Odyssey, are obviously at close quarters with their 
subject ; and if there is one thing certain about the civilisa- 
tion of the "Homeric Age" thus described, it is that it 
differs in nearly every important feature from that of the 
" Hellenic Age" of historical Greece. 

2. The Greeks, in fact, themselves regarded their earliest 
literature as antedating the chronological limits of their 
history, and already perceived that they belonged to a 
different order of things. In particular, the ethnography 
of the /Egean, preserved in an admittedly late and de- 
generate lay, differs uniformly from that of historic Hellas as 
far back as it can be traced, and those names are almost 


absent by which the Greek race was denoted historically ; 
by its western neighbours as "EAXiji'tc, by its eastern neigh- 
bours as 'laoveg (Javan). This inconsistency was attributed 
by the Greeks themselves to a period of invasion and 
migration analogous to that which broke up the Graeco- 
Roman civilisation of the Mediterranean. Dorian, 
Thessalian and Boeotian mountaineers were represented as 
forcing the barrier, or descending from the highlands, of the 
Balkans, bringing the old established " Achaean " civilisa- 
tion to an abrupt close, and reducing the /Egean, and 
mainland Greece in particular, to a chaotic and barbarous 
state, the recovery from which is the dawn of the historical 
Hellenic genius. 

3. Some facts within their own experience went to 
confirm this view. Here and there tribes retained the names 
and the mode of life of the earlier age ; or a noble family 
professed to trace its descent beyond the limits of current 
genealogy, and to identify itself with a Royal house of 
Achaean princes ; and here and there ruined fortresses 
remained, or ancient tombs had been disturbed, which 
seemed to confirm the description of Achaean splendour in 
the ballads. 

4. Thus much had been established from the beginning 
of Greek History onwards, and had not been seriously 
shaken by successive attempts to discredit the traditional 
view. The theories that the lays are comparatively late 
compositions, and that they stand in no close relation to 
a pre- Hellenic age ; that the Achaean Age is an invention, 
and the Period of the Migrations a hypothesis to explain its 
inconsistency with the facts of historical geography, all 
prove too much, and may be met with argument a ad 
hominem from the same traditional data. No literary 
critic of the Epic has yet purged himself of a sediment of 
traditional preconception ; and, in proportion as one or 
another has attempted to do so, he has been reduced to a 
merely agnostic position. 

5. Further, until very recent years, every attempt which 
was made to elucidate the civilisation of the Homeric Age 
by the monuments of early Greek civilisation rested upon 


the assumption that the representations of dress, armour, 
etc., of the sixth, fifth and fourth centuries B.C., were valid 
illustrations of poems which at the latest belonged to the 
seventh, and on an average were assigned to the ninth or 
tenth century. The reason of this was that Homeric sub- 
jects in Greek art are uniformly furnished with accessories 
of the age of the artist, and that until the study of Classical 
Antiquities began to be infected with the " evolutionary 
notions " which had already long been current in all other 
departments of Ethnography, the attention of students of 
Greek art and culture was strictly confined to mature and 
decadent art ; everything which could not be assigned to a 
century subsequent to the fifth was either dismissed as 
barbaric, or discounted as a " Phoenician importation " ; the 
part which " Phoenician " fables, ancient and modern, have 
played in the historical study of the Mediterranean area will 
be considered briefly later on. Such, for example, was the 
received opinion — so far as there was one — of such dis- 
coveries of pre-Hellenic culture as those of M. Fouque's 
expedition to the Island of Santorin (Thera, 1862), where, in 
the course of a geological investigation, a primitive settle- 
ment was found under a thick bed of volcanic debris, or of 
those of MM. Salzmann and Biliotti (1868-71), who in 
searching for antiquities in Rhodes found at Ialysos, for the 
British Museum, a magnificent collection of early vases 
which are now known to be Mykenaean, and second only in 
quality and variety to those from Mykense itself. The 
Santorin settlement was simply taken to confirm the legend 
of the Phoenician colony of Kadmos (Hdt. iv., 147), and 
the vases from Ialysos were explained as the barbarous but 
immediate predecessors of those from Kamiros, were classed 
with them as " Grseco-Phcenician," and were referred to the 
seventh century, in spite of the absence of Egyptian objects 
of the twenty-sixth Dynasty, and the presence of objects of 
the eighteenth : a view which in certain quarters is not 
yet quite extinct. 

6. It was not till 1871 that Dr. Heinrich Schliemann 
was enabled to execute his lifelong ambition of testing with 
the spade the Greek tradition that the site of the Grseco- 


Roman town of Ilion was also the site of Homer's Troy. 
The tradition had indeed been sorely handled by Deme- 
trios of Skepsis, a local antiquary of the second century 
B.C., on the geological ground that the Plain of Troy is of 
recent alluvial formation ; and by other critics on the score 
of inconsistency with the Homeric narrative. But the Bali 
Dagh, the site suggested by Demetrios, and in fact the 
only alternative, is far more inconsistent, and is put 
absolutely out of question by Dr. Schliemann's discoveries. 
In successive seasons (1S71-3, 1876-82) he laid bare not 
one, but six cities, built one after another on the same site, 
and forming an accumulation of walls and debris some 
thirty feet deep ; and, among these, two additional layers 
have been distinguished in the confirmatory excavations of 
Dr. Dorpfeld, 1892-94. The latter, however, indicate that 
Dr. Schliemann's earlier work was not, from the circum- 
stances of the case, sufficiently closely watched throughout, 
and that in some cases objects were probably picked up at 
lower levels than those to which they properly belong. In 
particular, it is not clear that the cache of jewellery and 
plate known as the "Great Treasure of Priam" was not 
hidden originally in a shaft of some depth. 

7. Dr. Schliemann claimed as the Homeric Troy the 
Second Town from the bottom, which had perished by fire, 
and in which the " Great Treasure " was found. But the 
Sixth Town, which Dr. Schliemann described as Lydian, 
was shown by Dr. Dorpfeld in 1892-93 to be larger and 
more important than was at first supposed, and to cor- 
respond closely with the remains found subsequently at 
Mykenae and elsewhere. 

8. With the same purpose in view of testing the 
Homeric tradition, Dr. Schliemann proceeded in 1875-6 to 
excavate the citadel of Mykenae, in the Peloponnese, the 
traditional centre of the Achaian feudal confederacy. Here 
the results were equally unexpected, but no less confirma- 
tory of the legend. A civilisation was brought to light 
wholly un- Hellenic, but far from barbarous ; greatly in 
advance of all but the latest layers of Hissarlik, and 
presenting already the marks of decadence after a protracted 


career. The pottery, the personal ornaments, and in fact 
the whole cycle of the art, were at once recognised as 
identical with those of Ialysos, while the stone-fenced 
burial-place discovered just within the " Lion Gate" of the 
citadel, with its six " shaft graves " and their enormous 
wealth of gold vessels and ornaments, seemed ample con- 
firmation of the legendary wealth of " golden Mykenae," 
and was proclaimed, in the first enthusiasm of the discovery, 
as the tomb of Agamemnon himself. The further re- 
searches which have been made almost continuously from 
1886 onwards by M. Tsountas for the Greek Archaeo- 
logical Society have confirmed in all essential points the 
first general impression, but the discovery of later tombs in 
the lower quarters of the town has made it possible to trace 
an order of progress and to extend the limits of the period. 
9. Subsequent excavations at Tiryns and Orchomenos 
by Dr. Schliemann, and on a number of other sites in 
Greece and the yEgean Islands by the Greek Archaeo- 
logical Society and the foreign Institutes in Athens, have 
demonstrated that this civilisation, which has acquired the 
provisional name of Mykenaean, is widely represented in 
the yEgean area and especially in its southern part ; that 
its influence extended over the Central and Eastern 
Mediterranean from Sicily to Cyprus; that it penetrated, 
intermittently at all events, into Egypt, where its appari- 
tion can be approximately dated, and whence it imported 
much, and borrowed somewhat, but without losing its own 
individuality ; and, most striking of all, that, after a long 
period of apparently continuous maturity, it falls into a 
sudden decadence ; leaving, to all appearance, just the same 
gap between itself and the first traces of Hellenic Art, 
as we have noted already, on the literary side, between 
the Homeric Age and the beginning of Hellenic His- 
tory. It should be further noted, however, that in the 
last few years many facts have come to light, especially in 
Attica, in Crete, and, most of all, in Cyprus, which seem to 
indicate how that gap may eventually be filled. It is from 
the pottery, almost without exception, that the leading 
indications have been derived. Fragments of baked clay 


are practically indestructible, even though the vessels which 
they composed have been shattered. Moreover, all the 
unrefined varieties of clay, and many even of the best 
levigated, present features by which their place of origin 
may be recognised. Consequently, in this material, 
modelling and decoration can be perpetuated as in no other 
way ; and, what is more important, the intrinsic worthless- 
ness of earthenware has often preserved it from the dis- 
placement and destruction which almost inevitably overtake 
objects of gold, bronze, and marble. The resulting pre- 
ponderance of ceramographic references in the bibliography 
which follows these notes must therefore be taken as 
indicating the character of the evidence which is most 
accessible, and of the method which has actually proved 
most fruitful : not that the pottery really took so large a 
place in primitive art as might be inferred from its actual 
abundance, and its scientific importance. 

10. Consequently the study of Early Man in the JEgean 
has entered within a few years on a new phase, and pre- 
sents the following problems: (1) To reconstruct in detail 
the history of the Mykenaean civilisation ; its origin, its charac- 
ter, range and influence, and its decline ; (2) to investigate the 
causes of that relapse into barbarism, which both literature 
and archaeology attest ; (3) to determine the ethnological 
position of the race, or races, who originated, maintained, 
and overthrew it, and their relationship with the historic 
inhabitants of the same area ; and (4) as a special study, to 
determine the relation in which the Hellenic traditions of 
the Achaean Age, and the lays in which they were preserved, 
stand to the civilisation which they certainly seem to com- 
memorate, and which owes its discovery simply to the 
application to them of a new method of criticism. 



1 1. Palaeolithic Man seems to have left no traces in the 
Levant comparable with those in North Europe, or with 
the plateau and upper-gravel flints of the Nile Valley. But 
the scarcity of evidence is partly due to the indifference of 


the natives to such objects, and to the almost complete 
diversion of trained research into more obvious and attrac- 
tive departments ; partly also to the comparative rarity, 
except in Egypt, both of workable flints and of the high- 
level gravels in which they are usually preserved. From 
Greece itself only one palaeolithic implement is recorded 
hitherto ; a flint celt from Megalopolis in Arkadia (Rev. 
Arch., xv., 1 6 ff). 

12. Neolithic Man, however, can be traced over the 
whole area. Masses of hard crystalline rock are frequent 
and accessible, and furnished implements of characteristic 
types ; short full-bodied celts, more or less markedly 
conical behind, and ground to a rather obtuse edge. Ob- 
sidian was largely exported from Melos and Thera to the 
neighbouring islands, and to the mainland of Greece, and 
was worked up at Korinth and on several sites in Attica. 
Jade of good quality was sent from Asia Minor outwards 
across the yEgean ; but it is not yet clear whether the 
source of the common green variety is in Asia Minor itself 
or further east : the jade implements become commoner 
eastwards, and the finest collection from anysingle neighbour- 
hood is that brought by Mr. D. G. Hogarth in 1894 from 
Aintab in N. Syria (Ashm. Mus., Oxford). 

13. Tombs of this stage of culture have not been found 
— or sought — in sufficient numbers to justify discussion or 
to contribute any facts of importance. The necropolis of 
Psemmetismeno in Cyprus, for example, contains besides 
typical early Bronze Age tombs a still more primitive class, 
in which the pottery is exceedingly rude, and the charac- 
teristic red-polished ware of the early Bronze Age is 
wanting ; but though bronze is absent, no stone implements 
are present. On the other hand the few tombs recorded 
as containing stone implements are brought down by their 
general character well within the Bronze Age. 

14. Exception must however be made in favour of the 
Nile Valley, for Professor Flinders Petrie in 1895 found, 
at Ballas and Nagada, both tombs and villages of an 
invading race, apparently Libyan, which had brought the 
art of flint working to unequalled proficiency, and remained 


almost ignorant of the copper which was already in fairly- 
common use under the Sixth Dynasty, which immediately 
preceded their irruption into Egypt. But the significance 
of this discovery and of our very limited knowledge of the 
Libyan people and their civilisation will be better discussed 
at a later stage. 

15. On the other hand, several Settlements of the 
Neolithic Age have been examined. Typical is the lowest 
town of Hissarlik, though it has actually yielded a few 
simple copper weapons. The implements are of local flint 
and imported obsidian, of green-stone and allied rocks from 
the interior of the Troad, and of jade ; some of the common 
green Anatolian, others of finer yellowish kinds {cf. the 
specimen in Ashm. Mus. attributed to Melos), and one 
small celt of the pure white variety which is not known 
to exist native except in China. 

16. The fortifications and house walls of the "First City" 
are of very rough unhewn rubble ; its pottery is of local 
fabric, made wholly without the use of the potter's wheel, 
and almost uniformly tinted black by a carbonaceous pig- 
ment, intentionally applied and accentuated in the burning ; 
many of the forms are closely allied to those of the neolithic 
and early bronze ages in Central Europe, and of the corre- 
sponding deposits of Greece and Cyprus. This lowest 
settlement is separated from the rest by a layer of natural 
soil, which represents an interval during which the site lay 
desolate ; it is therefore distinctly older than the succeeding 
cities. But the advanced and special technique of the 
Pottery of the First City, and the fact that, on Schliemann's 
authority, copper implements already occur, indicate the end 
rather than the beginning of the Neolithic stage ; and the 
Neolithic evidence from elsewhere is best summarised here, 
before going further in the series at Hissarlik. 

17. Settlements of similar character, but each with its 
own local peculiarities, occur (r) on an unexcavated site, 
commanding the Bosphorus as Hissarlik commands the 
Dardanelles. (2) On the " Kastri " near Achmet-aga in 
Eubcea, a low hill fortified with earthworks and approached 
by a hollow way, like the hill camps of the south of England. 


(3) Beside Dombrena near Thebes in Central Greece : the 
site has not been described, but neolithic implements are 
very frequent : among them is a potter's burnisher of white 
quartzite (Finlay Coll., 280. Athens). (4) On the Acro- 
polis of Athens many implements and vases were entirely 
confused by the levelling of the summit in the fifth century 
B.C. ; on the south side (in the space afterwards known as 
the UtXapyiKov) is a layer of neolithic pottery with obsidian 
flakes and a potter's burnisher, almost wholly destroyed 
by the recent excavations, and only preserved where it is 
left to support the fragmentary walls of the Mykenaean 
settlement. The material of the pottery is Ilissos mud, 
not the Kerameikos clay of the Kephissos valley. (5) 
Beyond the Ilissos, between Hymettos and the sea, the 
exact site is unknown, potsherds are common on the surface. 
The many stone heaps in this district seem to have been 
accumulated from off the fields on to barren spots ; two, 
opened south-east of Kara in 1895, were quite barren; a 
tumulus north-east of Kara, surreptitiously opened, con- 
tained a Mykenaean interment (Ashm. Mus.). (6) Primitive 
pottery is common on the west end of the cliff which runs 
along the coast from New Corinth nearly to the site of 

18. The "Second City" of Hissarlik has marked points of 
similarity with the first, but represents a decided advance, 
and has notable characteristics of its own. The walls, great 
and small, are of better masonry below, and of sun-dried 
brick above, with bonding courses and terminal uprights 
(antae) of timber ; the centre of the fortress is occupied by a 
" chief's house," consisting of three oblong buildings with 
portico entrances at one end in a courtyard entered by a 
covered gateway. The pottery is still of unlevigated clay, 
and mostly hand-made ; it is no longer blackened as before, 
but either left as it is, or covered with a red slip, which con- 
tinues to occur in the layers above ; new and characteristic 
forms appear, some peculiar, others again common to 
Central Europe, to the Greek islands or to Cyprus. 
Stone implements are still in common use, but copper and 
bronze begin to be frequent though they are still of simple 


types. But the pre-eminent feature of the Second Town is 
the discovery of more than one buried " Treasure " of gold 
and silver jewellery and vessels, the latter certainly of 
local manufacture, for the forms closely correspond with 
characteristic types of the pottery. 

19. The Second Town perished in a general conflagra- 
tion, and the Third, Fourth and Fifth Towns above it 
never attained to anything like its magnificence. They 
mark, however, a gradual advance of civilisation and form a 
transition, more and more rapid as it proceeds, towards the 
Sixth Town, a quite distinct and well-marked settlement of 
" Mykenaean " invaders, in which imported pottery, and 
native imitations of this, occur alongside of fully developed 
indigenous forms, which again recall in characteristic details 
many Central European types. This Sixth Town is the 
only one which can be even approximately dated chrono- 
logically ; it is certainly prior to 1000 B.C., and need not be 
later than 1 300 ; the Fifth and lower settlements must of 
course necessarily be older than this. 

20. It has been already hinted that the " Treasure of 
Priam " may belong to a period somewhat later than the 
Second Town, though not so late as the sixth or 
" Mykenaean " Town. Whether this be so or not, we 
have in the jewellery an early example, perhaps a prototype, 
of the characteristic gold work of the Mykenaean Age ; 
but if the " Treasure " is contemporary with the layer in 
which it was found, the time limit for the whole series at 
Hissarlik must probably be contracted downwards. In 
any case we must believe that the earliest civilisation of 
Hissarlik was not so wholly barbarous as appears at first 

21. Imported objects found at Hissarlik indicate a wide 
range of foreign connections. The fragments of porcelain 
point to Egypt ; the lapis lazuli axe from a neighbouring 
site, to Turkestan ; the silver vases probably to the eastern 
half of Asia Minor ; the types of the bronze implements 
alike to Cyprus and to the Danube Valley ; and the amber 
to the shores of the Baltic. This wide commerce does not, 
of course, imply direct intercourse, but, from its geographical 


position on the Hellespont, Hissarlik must have been a 
point of convergence for any trade between the East and 
Europe, and the catalogue of the allies of the Trojans in Iliad 
II., though it refers to a later period, ranges them (i) up 
the Hebros Valley into the Balkans, and along (2) the 
North and (3) the West coast of Asia Minor; i.e., along 
three well-known routes of early trade. 

22. The metallic objects of Hissarlik are of particular 
value as links between two principal copper-working areas, 
Cyprus and Central Europe. The latter really falls 
beyond our present view, but must be noted — mainly to be 
rejected — as a possible source of the early Mediterranean 

23. The use of copper in Cyprus goes back far beyond 
the point where it can be dated with any certainty, and 
everything goes to show that, while southwards, namely, 
in Egypt under the Fourth Dynasty, Cypriote types appear 
from the first side by side with others which are 
probably Sinaitic, northward the same types extend, past 
Hissarlik, into the Danube Valley, and are imitated and 
amplified into derivative forms throughout Central Europe ; 
returning, almost unrecognisable, into the Mediterranean 
area in the series from Spain, which is clearly not directly 
derivative, and may be of comparatively late origin. 

24. The obvious suggestion that Central Europe may 
have worked copper independently is met (1) by the com- 
parison of the secondary forms, — e.g., only in Cyprus can the 
actual synthesis of double-bladed axe heads, by welding 
two simple ones, be observed ; (2) by the fact that, along 
with the characteristic and indigenous metallurgy, the 
ceramic technique of Cyprus, with red hand-polished sur- 
face and incised ornament filled with white earth, can be 
traced across Asia Minor and into South-eastern Europe ; 
the red slip as far as Brus in Transylvania ; the ornament 
into the Mondsee of Lower Austria, and the pile-dwellings 
of Switzerland, becoming ever more mongrel and degenerate 
as it proceeds. 

25. It is important to note that at Hissarlik a return 
current is already evident ; the pottery and the metal im- 


plements reproduce European types as well as Cypriote, 
and this is confirmed, not only by traditional and 
ethnological considerations, but also by the occurrence, 
somewhat later, in the yEgean area, not only of frequent 
amber, but of characteristically Danubian types of bronze 

26. The Bronze Age civilisation of Cyprus is, thanks to 
repeated researches, far more continuously and completely 
known than any other part of the area. It was undoubtedly 
of very long duration, and certainly follows that of the 
Stone Age without change or break ; and it is no exaggera- 
tion to say that, until a period between the twelfth and the 
eighteenth Egyptian Dynasty, Cyprus was in all essential 
respects in advance, not only of the coasts of Asia Minor 
and the /Egean, but even of the coast of Syria and 

27. All the earliest weapons, whether in Cyprus or 
elsewhere, in Egypt, or the Levant, are of almost pure 
copper. Tempering is effected, not by alloying with zinc or 
tin, or, as in the Caucasus, with antimony from the natural 
double-sulphide ore, but by " under-poling " the copper so 
as to leave it hard and even brittle from the presence of 
copper oxide. The same applies to the Egyptian copper 
weapons of the fourth, fifth, and even sixth Dynasty ; but 
Egypt, though later on it has important connections with 
Cyprus, obtained its first copper from the mines of Sinai, 
and has a set of typical forms peculiar to itself. Cyprus, 
however, supplied the Syrian coast with copper weapons 
down at all events to the time of the eighteenth Dynasty. 
Stone implements are very rarely found in Cyprus, 
and it is possible that either the island was not reached 
much before the beginning of the Bronze Age, or that its 
wealth of copper was discovered at once, and superseded 
the stone age prematurely. In its earlier stages metallic 
implements are rare, and the pottery — always made by 
hand — is covered with a bright red glaze which was polished 
with a stone or bone rubber (horse teeth were commonly 
used), and ornamented, if at all, either by incised lines or 
by pellets of clay rudely modelled after plants, snakes and 


horned animals. In its earlier part, therefore, the civilisa- 
tion, so far as it is known, is peculiarly uniform in character, 
and displays no trace of foreign influence ; except only that 
the characteristic red-polished glaze of the pottery, already 
mentioned, is almost identical with that of the Neolithic 
Libyan people of Ballas-Nagada, and of their " Amorite " 
kinsfolk in South Palestine. Even here, however, there is 
no evidence at present of imitation on either side. The 
strong influence which Cyprus exercised, through its copper 
trade, over the neighbouring coastland is best illustrated 
by the discoveries of Dr. Bliss at Tell-el-Hesy, on the 
coast plain of Palestine (Philistia), some sixteen miles from 
Gaza. The site consists of an acropolis with eight "Cities " 
superimposed as at Hissarlik. The mass of the remains 
represent an indigenous "Amorite" civilisation of low type, 
related, according to Professor Flinders Petrie, to that 
of the Libyan invaders of Ballas-Nagada. But bronze appears 
from the bottom of the series upwards, and iron already in 
City Four, which with City Three appears to be contemporary 
with the eighteenth Dynasty and the Mykenaean Age. 
The bronze types are derivative, partly from Cyprus, partly 
from Egypt ; and Cypriote importations of the later painted 
fabrics occur in Cities Two and Three together with native 
imitations. The red-polished pot fabric of Tell-el-Hesy, 
however, belongs to the Amorite civilisation, and is not 
necessarily borrowed from that of Cyprus. 

28. In the latter half of the Bronze Age, Cyprus with 
characteristic conservatism fell for a while slightly behind 
its neighbours, and began to import ornaments and articles 
of luxury from Egypt and the Syrian and Cilician coasts. 
In this stage the red-polished ware tends to deteriorate in 
colour and finish ; the bronze weapons become more 
numerous, and contain a higher percentage of tin, and 
occasionally jewellery of coarse silver-lead, all of native make, 
is found in the more richly furnished tombs. Babylonian 
cylinders occur rarely as imports, with a multitude of charac- 
teristic native cylinders. Egyptian scarabs and porcelain 
beads are also found rarely ; and with these again a very 
common variety of coarse crumbly porcelain badly glazed 


with a very faint blue : the pigment was evidently difficult to 
obtain, and was used but sparingly by the native artist. 
But meanwhile the discovery of the art of ornamenting the 
natural surface of clay vessels with an encaustic umber pig- 
ment, wherever it may have originated, seems to appear 
in Cyprus (where umber is extensively worked) at least 
as early as anywhere else ; first in company with, but later 
almost wholly superseding, the older mode of incising linear 
ornaments on a prepared and polished surface. 

29. The simply painted pottery is followed, though not 
immediately, by several other fabrics which, though probably 
native to Cyprus, are represented in some quantity on 
Egyptian sites of the twelfth Dynasty and later dates, and 
also in equivalent layers in the stratified mound of Tell-el- 
Hesy, in the "Hittite" Sinjirli, and sporadically else- 
where ; one very characteristic variety, with dark body, 
white chalky slip, and black almost glossy paint, has been 
found even so far afield as the Island of Thera, the Acro- 
polis of Athens, and the " Sixth City" of Hissarlik. 

30. The specimen from Thera was found in company 
with vases of a distinct and local style ; some still with 
coloured surface and incised ornament, others with simple 
painted patterns. The forms, however, and the whole 
fabric, are quite distinct from those of Cyprus, and show a 
graceful freedom which is quite new; though they are clearly 
derivative from a ceramic of the Hissarlik type. Most 
important of all, the wholly geometrical and mainly linear 
ornament which has been hitherto universal is combined 
with or replaced by a thoroughly and vigorously natural- 
istic study of animal and vegetable forms, and, in combina- 
tion with the latter, spiral motives appear, hitherto unknown 
but destined to a long and eventful career. These naturalistic 
and curvilineardesigns are notonlyrepresentedon the pottery, 
but are also frescoed upon the plastered walls of the houses ; 
they may consequently be taken to be locally characteristic. 
The settlement at Thera was found beneath a thick bed of 
volcanic debris, and had evidently been suddenly abandoned ; 
metallic objects are rare, but this may well be due, as M. 
Tsountas suggests, to the flight of the inhabitants — for no 


skeletons were found ; and a few copper implements and 
gold ornaments remained to confirm the inference from the 
pottery as to its position in the series. 

31. Settlements and tombs of the same character have 
since been noted in many islands of the Archipelago, especi- 
ally in Syros, Melos, Antiparos and Amorgos ; and this 
" Cycladic " type of ornament and general civilisation is not 
only closely paralleled by the earliest remains at Mykense, 
Tiryns, Athens and elsewhere, but is connected by an 
almost continuous series with the fully developed art and 
civilisation of the Mykensean Age itself. 

32. It should be noted that though Cyprus appears to 
have exported its own manufactures to the yEgean during 
this period, it was not in a position to influence or direct 
the Cycladic culture. But still less is there any trace that 
the younger and more vivacious school reacted at all upon 
the elder ; this was reserved for the full-grown culture of 

2,7,. It is at this period that the Cretan evidence, though 
as yet miserably incomplete, becomes of crucial importance. 
Crete shares, to begin with, the early bronze age civilisa- 
tion of Hissarlik and Cyprus, resembling the latter more 
closely ; but it is not till the Cycladic stage is reached that 
we have more than the most fragmentary evidence. In the 
Cycladic period and in the succeeding age Crete was almost 
literally tKaro^woXiQ, the " island of an hundred cities," and 
certainly exercised a vigorous and continuous, perhaps even 
a predominant influence upon /Egean civilisation. At this 
point the wealth and variety of Cretan decorative art become 
conspicuous, and a chronological point of the very first im- 
portance and a clue to the origin of some characteristic 
motives are given by the recent demonstration of a frequent 
and fertile intercourse with Egypt in the time of the twelfth 
Dynasty. On the one hand, a very peculiar and local fabric 
of pottery from Kamarais in Crete has been found in twelfth 
Dynasty layers at Kahun ; on the other, the Cretan types 
of bronze implements are typically Egyptian, and twelfth 
Dynasty scarabs were not only frequently imported, but 
commonly imitated. In fact it is very probably from this 


quarter that the spiral motives, which are dominant in the 
Egyptian Art of the twelfth Dynasty, were introduced into 
the decorative repertory of /Egean art. 

34. The seal-stones engraved with Egyptian and deriva- 
tive spirals are closely associated in Crete with others 
bearing groups of symbols, more than eighty of which have 
been recorded, and shown to be hieroglyphic, by Mr. A. J. 
Evans. They exist in two series, of which the earlier is fully 
pictorial and naturalistic, the later conventionally abbre- 
viated into linear forms. Some of the former are closely 
analogous to certain Egyptian, others to certain " Hittite " 
hieroglyphs from Kappadokian monuments ; many of 
the latter are identical with graffiti on twelfth-eighteenth 
Dynasty pottery from Kahun, Tell-el-Hesy and elsewhere, 
and some are probably prototypes of symbols which per- 
sisted in the Phoenician, Greek and Lykian alphabets, and 
in the Cypriote syllabary. This hieroglyphic system is not 
confined to Crete, though it is far best represented there 
as yet ; the pictorial seal-stones are distributed over the 
Cycladic area ; and two inscriptions in the linear character 
have been found on vases at Mykenae. Dr. Kluge, of 
Magdeburg, believes that he can translate these hiero- 
glyphic inscriptions into a dialect of Greek. 

35. We now come to what is, even literally, the Golden 
Age of the early Mediterranean cycle. " Mykensean " Art 
is still best and most completely illustrated by the long 
series of discoveries in the plain of Argos, which at once 
revealed its existence, and have given to it a name. The 
monuments and the civilisation of Mykenae and Tiryns 
have been repeatedly, though never yet really adequately, 
described, and have given rise to the most divergent 
theories as to their date, their origin, and their relations 
with what precedes and follows them. The following 
points are those which are chiefly made clear by the most 
recent researches. 

36. The limits within which Mykensean sites are dis- 
tributed may now be defined with some approach to 
accuracy, and no less the wider area over which Mykenaean 
civilisation had a living influence. With the exception of 



the "Sixth City" of Hissarlik no Mykenaean settlement is 
known on the mainland of Asia Minor. Isolated vases are 
reported from Pitane in JEoYis, from Mylasa in Karia, and 
from Telmessos in Lykia, and the early necropolis of 
Termera (Assarlik) near Halikarnassos (Budrum), though 
of distinctly indigenous character, is strongly influenced, at 
the very end of the period, by late Mykenaean models from 
the neighbouring islands. Among the latter, besides the 
great settlement at Ialysos in Rhodes, every island appears 
to be represented from Rhodes southwards to Crete, and 
northwards as far as Patmos. Both in Melos and in Thera 
Mykenaean settlements are found distinctly superimposed 
on the Cycladic already mentioned, and others are indicated 
by isolated finds throughout the Archipelago. On the 
mainland of Greece, Lakonia is represented by two sites 
Kampos and Vaphio (Amyhlae), the latter with a princely 
"beehive tomb" like those of Mykenae ; Argolis by 
Mykense, the Heraion temple-site, Tiryns, Nauplia, 
Trcezen, Epidauros, and the islands Kalauria and ^gina ; 
Attica by Athens, Eleusis, Acharnae (Menidi), Aliki, Kara, 
Spata, and Thorikos ; the rest of Central Greece by 
Megara, Antikyra, Thebes, Tanagra, Levadia, Orchomenos 
and several smaller sites in the Kopais marshes; North 
Greece by Pagasae (Dimini near Volo) in Thessaly. 

$*]. In the West there are no Mykensean settlements 
known further than Kephallenia and Ithaka; but Mykensean 
vases occur in domed rock tombs at Syracuse, and there is 
much indirect evidence of Mykenaean influence on the later 
Bronze Age style in Sicily and South Italy. Further than 
this, it is clear that on the Adriatic coast of Italy Mykenaean 
imports and models determined the character of the later 
Bronze Age, and that in the transition from Bronze to 
Iron at Hallstatt in the Tyrol, a definitely Mykenaean strain 
can be detected. But in both these cases the contact is 
with later and already quite decadent types, such as are re- 
presented in the Lower Town of Mykenae ; in particular 
fibulae are always present, and of these the secondary and 
distinctly Sub-Mykenaean types are only very rarely absent. 

38. Eastwards, Mykensean imports are found frequently 


in Cyprus, in the latest class of Bronze Age tombs, 
and give a very distinct character to the necropoleis 
of Episkopi (Kurion), Enkomi (Salamis), Pyla, Niko- 
lidhes, and Laksha-tu-Riu. Native imitations increase in 
frequency, and eventually supersede the importations and 
fix the leading features of the art of the early Iron 
Age, e.g., at Kuklia (Paphos), Lapathos and Katydata- 
Linu. In Egypt again, Mykenaean importations are found 
in great quantity, associated with the later Cypriote fabrics 
and stimulating copious native imitation in layers of the 
eighteenth Dynasty at Illahun, Gurob, Tell-el-Amarna. 
These last finds confirm the date already inferred from 
the occurrence of eighteenth Dynasty scarabs and porcelain 
ornaments at Ialysos and at Mykenae, and fix the general 
chronology of the Mykenaean Age beyond all question. The 
contrary opinion, that the Mykenaean civilisation immediately 
precedes the Orientalising culture of the seventh-sixth 
centuries, and consequently itself descends as late as the 
eighth-seventh centuries, has been vigorously urged by a 
few English students, but has long been abandoned by all 
who have had first-hand experience of the conditions of 
discovery. The premature contention that the fortress of 
Tiryns was Byzantine deserves mention, but is obsolete. 

39. It is in Egypt also, moreover, that the first notice 
occurs of the actual peoples who transmitted the civilisation 
in question, and this in a peculiarly suggestive connection. 
In the fifth year of Merenptah (1225) and under Rameses 
III. (1 1 80- 1 150) the western frontier of Egypt was seriously 
threatened by a Mediterranean coalition, of which the 
Libyans were the principal members, but which included 
under the general description of " the peoples of the isles 
of the sea " a number of tribes whose names, though much 
distorted in the Egyptian hieroglyphic records, strongly 
resemble those of Achaians, Danaans, Ionians, Teucrians, 
Tuscans or Tyrrhenians, and perhaps Sicilians and 
Sardinians. Neither these names, of course, nor yet the 
apparent resemblance of their arms and furniture, as depicted 
in Egyptian paintings, can give more than a plausible pre- 
sumption of identity either with historical /Egean races or 


with the representatives of Mykenaean civilisation. But the 
analogies are on all sides so close, that the identification is 
usually accepted, and that as soon as even the outlines of 
the history and civilisation of Libya during the Bronze Age 
are ascertained, we shall be in a position to formulate 
the real relations which then existed between Libya 
and the /Egean, and probably also to trace more clearly to 
its source the very remarkable realistic instinct which dis- 
tinguishes the art of the y^Egean from all contemporary 

40. The sudden collapse of the Mykensean civilisation, 
which was indicated to begin with, is roughly coincident with 
the first appearance of Iron in common use in the Levant, and 
the attempt has been made, though on no direct evidence, 
to connect the two tendencies. All the facts go to indicate 
that, so far as the Mediterranean area is concerned at all 
events, iron makes its appearance first on the Syrian coast, 
in the period which immediately succeeds the downfall of 
Egyptian suzerainty in that area under the nineteenth and 
twentieth Dynasties: e.g., at Tell-el-Hesy iron occurs down to 
the fourth "City" (= eighteenth Dynasty). The ambiguity 
of the Egyptian allusions under the eighteenth and previous 
Dynasties makes any earlier date uncertain, and iron has 
not been actually found in Egypt before the twenty-sixth 
Dynasty, 650 B.C. In Cyprus, where the evidence is com- 
pletest, and where abundant native ores have certainly been 
worked from an early period, iron suddenly becomes very 
common just at the point when Mykensean vases are ceasing 
to be imported, but when, on the other hand, Mykenaean 
conventions have already begun to influence profoundly the 
native scheme of ornament. At Mykenae itself iron occurs 
first as a " precious metal " and in the form of signet rings, at 
the stage where decadence begins to be rapid, but it is not 
put to practical uses till the moment where the series breaks 
off, and the same is the case in other Mykenaean sites in 
the iEgean ; one iron sword was found in the Vaphio " bee- 
hive ". 

41. Up the Adriatic again it is with the early fibulae and 
quite degenerate Mykenaean art, that iron makes its appear- 


ance, at Novilara ; and at Hallstadt ; and here again, both in 
tradition and among the finds, there is evidence that the 
metal became established first as an ornamental rarity, and 
only subsequently as a substitute for bronze. 

42. But though in its principal centres Mykensean 
civilisation has all the appearance of having been suddenly 
and violently extinguished, this must not be taken to be 
universally the case. In Argolis (at Tiryns, and the Heraion), 
in Attica, and in Melos, for example, there is every reason to 
believe that the Mykenaean civilisation survives, though in 
very degenerate phases, into the period when Iron and the 
characteristic art of the early Iron Age are already well 
established ; and at Nauplia and the Attic Salamis, and 
still more in Crete, in Karia, and in Cyprus, the stages may 
be clearly traced by which, so far as in it lay, the Iron Age 
took up its inheritance from the Age of Bronze. The 
nature and the result of this transference are easily sum- 

43. It has been already indicated, firstly, that through- 
out the Eastern Mediterranean, in fact throughout the whole 
range of the Mediterranean Early Bronze Culture, the 
indigenous system of decoration is instinctively rectilinear 
and geometrical ; secondly, that in the Cycladic area and 
in the Middle Bronze Age a quite irreconcilable and purely 
naturalistic and quite heterogeneous impulse appears ; and 
thirdly, that the fully formed Mykenaean style, when it 
appears, is, in spite of its far superior technical skill and 
elegance, already beginning to stagnate in many depart- 
ments ; the gem-engraving and modelling developing last, 
and retaining their vigour and elasticity latest ; whereas 
the ceramic decoration, which appears in its noblest 
form at Thera and at Kamarais, is the first to exhibit the 
conventional and mechanical repetition of a shrinking 
assortment of motives. We may now add, fourthly, 
that this failure of originality permitted of a recrudescence 
of the rectilinear instinct which, though overwhelmed for 
the time by the naturalistic and curvilinear principles, had 
co-existed with them throughout ; and that both floral and 
spiral motives, once allowed to repeat themselves without 


reference to their models, are transformed automatically 
into the latticed triangles and maeanders, which are the 
commonplaces of rectilinear design. 

44. At this point the survey must close, for now, on 
geometrically engraved tripods, and geometrically painted 
vases, appear Hellenic inscriptions in alphabetic characters. 
Borrowed Oriental, and especially Assyrianising, motives 
intrude themselves into the panels of the rectilinear orna- 
ment, and attempts are made, however ineffectual, to 
represent first animal and then human forms. Now, in the 
development upward out of the " Dark Age," Hellenic 
history begins to reckon onward from the Trojan Era and 
from Olympic and kindred lists ; and Hellenic art no longer 
forward from the eighteenth, but backward from the twenty- 
sixth Dynasty. 



N.B. The references which follow are grouped under 
the numbers of the paragraphs of the text. They only 
indicate the primary researches and theories, and must be 
compared with the fairly full references in Perrot and 
Chipiez, Histoire de I Art. VI, La Grece Prehistorique, 
1895, an d with the current notices of discoveries scattered 
throughout M. Salomon Reinach's invaluable " Chroniques 
d'Orient " published in the Revtie ArchcEologique, of which 
the years 1883- 1890 have been republished separately 
(Paris, Firmin Didot, 1891). 

6. Dr. Schliemann's Researches. 

Schliemann. Ilios. (German and Englished.), 1881, (French 
ed., including " Troja"), 1885, (German and English), 1884. 
Atlas Troj. A Iterthumer (photographs), 1874. 
Mycence ,, ,, 1878. 

Ithaka, etc., ,, ,, 1879. 

OrcJwmenos ,, ,, 1881. 

Tiryns ,, ,, 1886. 

SCHUCHHARDT. ScJiliemanii s Excavations (German, Leipzig, 
1890); E. T. Macmillan, 1891. 


11. The Stone Age. Sp. LAMBROS. 'Ia-opiKa MeXeryj/xara 

{Historical Essays), ch. i. 
Dumont. Materiaux pour servir a Fhistoire primitive de 

rhomme, 1872 (Finlay Collection). Revue ArchcEologique, xv., 

pp. 16-19, 356 ff., xvi., p. 359 (1867). 
PAPPADOPOULOS. AiQwi) kiroyy) ev rfj Mifcpa ' Aa'ia (Stone 

Age in Asia Minor), Smyrna, 1875. Cf. Bulletin des 

Correspondances Helleniques, ii., p. 8, 1876. 
FlNLAY. UapaTtipriaei<i {Observations), Athens, 1869. 

13. Psemmatismeno. DtJMMLER. Athenische Mittheilungen, xi., 

pp. 214-6, 1886. 
Bronze Age Tombs with Neolithic Implements. At Kurion in 
Cyprus. Archives des Missions, xvii., p. 6. Cypr. Museum 
Catalogue, No. 470 (Oxford, 1 896). At Tiberiopolis in Phrygia. 
J. A. R. Munro. Journ. Roy. Geog. Soc. (forthcoming). 

14. Ballas-Nagada. Catalogue of Exhibits, University College, 

London, July, 1895; Academy, 20th April, 16th July, 
1895 (Report forthcoming). 

15. Jade. SCHLIEMANN (Maskelyne). Ilios. (English), p. 240. 
FISCHER. Neplirite u. Jadcite . . . uach Hirer Urgesch. u. 

Ethnogr. Bedeutung, Stuttgart, 1875. 
Davies. Geol. Mag., second decade, v., 4, April, 1878. 

16. Hissarlik, v. § 6, SCHLIEMANN. 

NORMAND. La Troie d'Homere (popular, well illustrated). 

17. Thymbra. SCHLIEMANN. Ilios. s. v. 

Boz-oyuk (Phrygia). Jahrbuch d. K. Akademie, Berlin, xi., 

1896. Anzeiger, p. 34. 
Salonika. Jahrbuch, I.e. 
Thessaly. Mitth. Ath., p. 99 ff., 1884. 

18. Bceotia. Jahrbuch, 1895. Anzeiger, p. 32. 

J.H.S., pp. 54-56, figs. 10-13, 1884. 
Attica. Mitth. Ath., p. 138, fig. 31, 1S93. 
Jahrbuch, p. 16, 1893. 
22ff. MUCH. Die Kupferzeit in Europa (second edition), Jena, 1893. 
Naue. Die Bronzezeit auf Cypern. Korresp. Blatt, p. 124, 1888. 
VlRCHOW. Zeit. d. Deutsch. Gesellsch. d. Anthrop., xii., 73. 

27. Copper and Early Bronze with but little Tin. J. H. GLADSTONE. 

Proc. Brit. Ass. (Nottingham), p. 715, 1893. Trans. Soc. 
Bibl. Archeology, xii., pp. 227-234. Flinders Petrie. 
Zeitschr. f. Ethu., p. [477], 1891. BLISS, I.e. 
Tell-el-Hesy. BLISS. A Mound of Many Cities, 1894. 

28. Cyprus. Sandwith. Archcsologia, 1877. 
DtJMMLER. Mitt. Ath., xi., 1886. 
OHNEFALSCH-RlCHTER. Kypros the Bible and Homer, 1892. 


Myres and OHNEFALSCH-RiCHTER. Cyprus Museum Catalogue, 
Oxford. 1896 (in the press). 

30. Thera. FOUQUE. Santorin. Archives des Missions, ser. 2, 

vol. iv. 

31. Cyclades. Dummler. Mitth. Ath., xi., 1886. 
Antiparos. Bent. /. H. S., x., 1887. 

33. Crete. A. J. Evans. Journ. Hellenic Studies, xiv., pp. 276-372, 

1894 (republ. "Cretan Pictographs," etc., Quaritch, 1895). 

34. y£gean Hieroglyphic System. Evans, I.e. 
KLUGE. Magdeburger Zeitung, 1896. 

35. Mykenaean Civilisation in general. v. Bibliogr. in PERROT, 

vi., q.v. 
TSOUNTAS. Mvxrjvai fcal Muk. ttoXlthtpos (Mykenae and 

Myk. Civilisation), Athens, 1893. 
TSOUNTAS. s E^>rffi€pU 'ApxaioXoyL/cr) {Journal of Gk. Arch. 

Soc), 1 886- 1 894, passim. 
PERROT and Chipiez. Histoire de I'Art, vi. (la Grece 

Prehistorique) (E. T.), 1895. 
FURTW/ENGLER u. LcESCHKE. Myk. TJiongefdsse, 1879. 
FURTW^ENGLER U. LcESCHKE. Myk. Vasen, 1886. 
POTTIER. Vases Antiques du Louvre, I., p. 181 ff., 1896. 
HELBIG. La Question Myce'nienne. Paris, 1896. 
^6. Mykenaean Sites, Asia Minor : — 

Hissarlik, "VI." DCERPFELD. Troja, 1893. Mykenische 
Vasen, p. n. Reinach. Rev. Arch., 1893, i., p. 357. Rev. 

Arch., 1895, i., p. 1 13. 
Pitane (zEolis). PERROT, vi., Fig. 489-91. 
Lemnos. Rev. Arch., xxvii., 1895, Nov. -Dec. ; cf. Smyrna 

Telmessos. Mitth. Ath., xii., pp. 228-230. 
Thessaly. WOLTERS. Mitth. Ath., xiii, p. 262, PI. viii.-xi., 1889. 
Orchomenos and neighbourhood. SCHLIEMANN, q.v. De 

Ridder. B. C. H., p. 137 ff., 1895. Esp. Gha. De Ridder. 

B. C.Lf.,p. 271 ff., 1894. Noack. Mitth. Ath,, xix., 1894. 
Daulis. {Athens: National Museum), unpublished. 
Antikyra (Phokis). Lolling. Wolters' Mitth. Ath., xiii., 

p. 267, 1889 (identified with Medeon). 
Athens. TSOUNTAS. *E<f>. 'Apx-, 1891, p. 27 ff. 

GR/EF. Jahrbuch, 1892. Anzeiger, p. 16 ff. 
Wide. 'Adtjvcuop, ii., 1895, 168. 
Eleusis. Philios. 'Ecp. 'Apx-, 1889, p. 171. 
Koropi. Bruckner. Mitth. Ath., xvi., p. 200 ff., 18 

(identified with Pallene). 
Sal am is. (Athens: National Museum.) 


/Egina. Evans. /. H. S., p. 195 ff., 1892-93 (Gold-find). 

Reinach. Rev. Arch., November-December, 1895. 
Kalaureia. Wide. Mitth. At//., xx., p. 297, 1895. 
Troezen. Reinach. Chroniques, p. 628. 

Epidauros. [Athens: National Museum.) 'Apx AeXrlov, 1888. 
Kephallenia. Wolters. Mitth. Ath., xix., pp. 486-490. 
Crete. Milchhcefer. Die Anfange d. Kunst, p. 122 ff. 
Evans. /. H. S., xiii., pp. 276-372, 1894 (republ. "Cretan 

Pictographs," Ouaritch, 1895). 
FURTW.ENGLER u. LCESCHKE. Myk. Vaseil, pp. 22-4. 
Halbherr and ORSI. Museo Italiano, II., p. 908, pi. xiii.-xiv. 
Haussoullier. B. C. H., 1880, pp. 124-7. 
Joubin. B. C. H., 1892, p. 295. 
ORSI. Monumenti AnticJii d. Accad. d. Lincei, I., p. 201 ff., 

Perrot, vi., p. 451 ff. (bibliography). 
Sicily. ORSI. Bulletino di Paletnologia Italiana, xviii., pp. 

193 ff., 206 ff , xx., p. 257 ff. Necropoli Sieu/a, p. 30 ff. 
Spain. GASCON de GOLOS. Saragoza, i., pi. iii., p. 40. 
39. Chronology — For eighteenth Dynasty dates : — 

Flinders Petrie. /. H. S., xii., pp. 199-205, 1891. 
Perrot, vi., p. 1000 ff. 
For later dates (summary) : — 
REINACH. Chroniques, pp. 420, 575 ff Rev. Arch., p. 75, 1893. 

Classical Review, p. 462 ff, 1892. Times, 6th January, 1896. 

Academy, nth January, cf. 1st February, 1896. 
Torr. Memphis and Mykence. 1896. 
For " Byzantine " Tiryns (summary) : — 

Reinach. Chroniques, p. 280 ff, V Anthropologic, p. 701, 1893. 

J. L. Myres. 


THERE is, perhaps, no better illustration in geology of 
the value of detailed work than that which is fur- 
nished by the group of organisms, to the consideration of 
which this article is devoted. Formerly viewed with sus- 
picion by biologist and geologist alike, and frequently 
altogether ignored, we find the graptolites now treated 
with respect even by those who have not devoted special 
attention to them. Their value is generally recognised as 
aids in the determination of the age of strata, but besides 
this, a detailed study of the group will undoubtedly throw 
light upon the physical and climatic conditions under which 
the strata containing graptolite remains were deposited, and 
also upon the evolution of the various forms of graptolites. 
Every one will admit that the appreciation in which grapto- 
lites are now held is largely due to three papers by Professor 
Lapworth, one of which treats of these organisms from a 
biological (i), and the second (2) and third (3) from a 
stratigraphical point of view ; and the publication of these 
papers is doubtless largely responsible for the appearance 
of a large number of memoirs devoted to a study of the 
group under consideration which have been written of 
recent years. These recent memoirs it is the object of 
this paper to consider. 

The memoirs, early and more recent, treating of the 
graptolites are scattered through a variety of publications, 
but an excellent bibliography compiled by Otto Herrmann 
and published in his Inaugural Dissertation (4) gives a list 
of these memoirs up to and including the year 1883. Even 
with this guide the student has much difficulty in obtaining 
access to some of the publications, and a general monograph 
of the graptolites has yet to be written. In the list of 
" Monographs which are promised or are in course of 
publication " appended to the last " Monograph of the 
Palaeontographical Society" we note "The Graptolites," by 
Professor Lapworth, and all geologists must hope that ere 


long the professor will give to the world the full results of 
his prolonged researches into the history of the group. 
This monograph must necessarily be confined to an account 
of the British graptolites, but when that is complete surely 
Professor Lapworth will treat of those of other countries 

The graptolites, at one time referred by some writers to 
the Hydrozoa, by others to the Polyzoa, are now generally 
admitted to belong to the former class, though the exact 
value of the sub-division is not definitely settled, for whereas 
we find Professor von Zittel in his Paleontology treating of 
them as a sub-order, Graptolithidse (= Rhabdophora, All- 
man), divided into the groups Graptolitoidea Lapw. and 
Retioloidea Lapw., Nicholson and Lydekker {Manual of 
Paleontology) place them in a sub-class (Graptolitoidea). 
In these works the general structure of the graptolites is 
described, though, as will be seen in the sequel, one structure 
supposed to be absolutely characteristic of all graptolites, 
namely the virgula, is not really so. Comparatively little 
has been added to the knowledge of the histology of the 
graptolitoidea furnished by H. Richter (5), though some of 
his results have been confirmed by Professor Sollas (6) ; 
and additional information has been supplied by Professor 
S. L. Tornquist (7) and Dr. Perner (8). Some of the most 
important papers published of recent years treat especially 
of the mode of growth of the proximal portions of the 
graptolites. The first of these by Tornquist (9) is occupied 
with a description of sections through several deprionidian 
graptolites. The author distinguishes the obverse from the 
reverse aspect of the polypary, and also introduces two 
terms to distinguish its right and left portions — the " primor- 
dial " portion, containing the "primordial" series of hydro- 
thecae, is marked by the possession of the earliest hydro- 
theca, whilst the other portion is termed the "second" 
portion and possesses the second series of hydrothecse. 
When the obverse aspect of the polypary is turned towards 
the observer the primordial series of hydrothecae is in- 
variably on the left hand. The sicula sends out what the 
author terms a " connecting canal ' which opens into a 


" biserial chamber," thus producing a connection between 
the various parts of the polypary. These features are 
common to all the forms described by the author, but the 
forms differ in other respects. In Climacograptus scalaris 
Linn, and Climacograptus internexus Tornq. the biserial 
chamber communicates with two uniserial canals separated 
from one another by a median septum. In Diptograptus 
palmeus Barr. the septum scarcely extends through half the 
thickness of the polypary, whilst in Cephalograptus cometa 
Gein. it is " reduced to a narrow 7 fold of the obverse peri- 
derm," and in Diptograptus bellutus Tornq. it is altogether 

Two papers by Wiman (10) treat of the structure of the 
Diptograptidce and of Monograptus. Notices of these papers 
by E. M. R. Wood and G. L. Elles appear in the Geological 
Magazine for 1895, p. 431. The accounts of the structure 
of the sicula, and of those parts of the polypary immediately 
in contact with it, are largely confirmed by Holm in a paper 
to be noticed immediately, but the statement that the Dip- 
tograptida; are monoprionidian because the sicula gives rise 
to only one bud (which is on the right hand side) involves 
a special use of the term monoprionidian which will hardly 
meet with general acceptance. 

A most important paper by Gerhard Holm must now be 
noticed (11). Holm has had the advantage of studying 
some beautiful material derived from the J^aginatus-Yimestone 
(of Areing age) from various localities in the northern part 
of the Island of Oland ; the graptolites of this limestone he 
has succeeded in freeing from the matrix, thus rendering 
them serviceable for detailed study. (The method of re- 
moving the matrix is described by Holm in an article in 
Bihang K. Vetensk. Akad. HandL, Bd. xvi., 1890.) In the 
present paper he gives reasons for supposing "that the 
earlier development of the proximal part — the first three 
thecae — in all the bilateral or diprionidian forms of graptolites 
is in the main the same, and has taken place through the 
formation of only one bud on one side of the sicula — -or first 
theca, as I believe it is — which side is always the same in 
relation to the later development of the polypary. From 


this bud thereafter is developed partly the second theca, 
partly the canal — ' connecting canal ' — which connects both 
halves of the polypary, and which in the first place gives 
origin to the third theca (= first theca on opposite side of 
sicula), and partly also the common canal which connects 
the second theca with the succeeding ones." He describes 
the " sicula " which consists of two distinct portions, the 
"initial part" which he believes to correspond with the 
original " chitinous covering of the free zooid germ or em- 
bryo," and the apertural part which has the same function 
as a theca and may therefore be justly considered as the 
first theca. Accordingly Holm's second theca corresponds 
to Tornquist's primordial one, and his third to Tornquist's 

The sicula in the bilateral graptolites does not occupy a 
central position, being partly embraced on one side by the 
connecting canal, whilst on the other side it is more or less 
superficial. The sicula side is termed the "anterior," and 
the other the "posterior". These are used in the same 
sense as that in which Tornquist employs the terms "ob- 
verse aspect" and "reverse aspect". The author gives a 
full account of the connection between the sicula, the first 
theca, the first bud, from which " arises almost simul- 
taneously with the left theca the common canal for the 
left half of the polypary, and the connecting canal which 
crosses the dorsal side of the sicula and gives origin to the 
third (or, better, the right) theca lying on the right side of 
the polypary, and also the common canal for the right side 
of the polypary," and describes the growth of these in 
Didymograptus minutus Tornq., D. gracilis Tornq. mut., D. 
gibberulus Nich., Tetrgraptus Bigsbyi Hall, and Phyllo- 
graptus angustifo/ius Hall. 

He maintains that a virgula cannot occur in any 
graptolites of the families Dickograptida, Dictyogr apt idee, 
and Nemagraptidce, or in the genus Dicellograptus of the 
family Dic7'anogi r aptid&. The true virgula commences 
near the apex of the sicula as a prolongation of the same, 
and corresponds with the thread-like prolongation of the 
sicula which has long been known in Didymograptus 


gibberulus, and certainly occurs in many other forms of 
Dichograptidce. Another filiform appendage which might 
be spoken of as the false virgula " originates as a result of 
growth within the apertural end of the sicula at some 
distance from the initial portion. This later structure 
stands evidently in no relation whatever to the real 
virgula, but may be regarded as an apertural spine." The 
significance of these filiform processes has not yet been 
fully explained, but the possession of a true virgula must in 
future be omitted from diagnoses of the characters of the sub- 
class or sub-order of the graptolites. Holm's researches 
fully confirm Tullberg's inference that Phyllograptus belongs 
to the family Dichograptidce, and the family Phyllograptidcz 
must now be abandoned. Another interesting point bear- 
ing upon classification is the position from which the bud 
grows out of the sicula. " In Phyllograptiis it is situated 
quite close to the apex of the sicula, in Tetragraptus 
Bigs by i Hall probably slightly lower down, in Didymograptus 
miniUus Tornq. somewhat below the middle of the 
sicula, in Didymograptus gracilis Tornq. Mut. still nearer 
the aperture ; but in Didymograptus gibberulus Nich. the 
position is almost the same as in Pliyllograptus." The 
reference of the genus Azygograptus to the Nemagraptidce 
on account of the stipe being developed from the central 
part of the sicula on one side is therefore unnecessary, and 
the general characters of Azygograptus leave no doubt 
that it belongs to the Dichograptidce ; indeed Holm in the 
paper under consideration describes a form which is possibly 
intermediate between Didymograptus and Azygograptus. 

The association of a number of graptolites of the same 
species in a fairly symmetrical manner has long been 
known. James Hall in plate xiv. of his classic work on 
graptolites (12) figures a diprionidian graptolite under the 
name of Retiograptus teutaculatus, and in figure 9 is "an 
illustration of a compound form of the genus," possessing 
nearly twenty diprionidian stipes diverging from a common 
centre. James Dairon (13) also figures specimens of 
Monograptus occurring in partly symmetrical tufts, and 
remarks : " I am now thoroughly convinced that many, if 


not all, of the specimens of Monograptus may have been 
fixed to the sea-bottom, or to objects lying or growing on 
it, and not have been free-floating organisms, as has hither- 
to been supposed, until at last they were separated from 
their points of attachment by breakage or some other 
natural cause ". Recently a remarkable description has 
appeared (14) giving an account of specimens of Dipto- 
graptus pristis Hall and D. pristiniformis Hall from 
the Utica Slates. In these specimens the stipes occur in 
"compound colonial stocks which appear in the fossil state 
in stellate groups ". From observations on the specimens, 
the author infers "that the colonial stock was carried by a 
large air-bladder, to the underside of which was attached 
the funicle. The latter was enclosed in the central disc, 
and this was surrounded by a verticil of vesicles, the 
gonangia, which produced the siculae. Below the verticil 
of gonangia and suspended from the funicle was the tuft 
of stipes," the latter being so arranged that the " sicula- 
bearing end of the single stipes appears in the compound 
colonial stock as the distal one ". The paper is only an 
abstract of one which is promised shortly, and geologists 
will await with interest a full account of these remarkable 
specimens. The structure described as a funicle can hardly 
be looked upon as the analogue of the " organ" described 
by Hall under that name (which by the way has been 
proved by Brogger and Holm to be celluliferous in many 
species, so that Holm is doubtless correct when he says 
that a funicle has not been found in any graptolite). It 
is remarkable that the author should explain what he 
means by the assertion that the chitinous capsule which 
encloses the " funicle ' : on the specimens described is 
identical with the "central disc ' : of the compound 
fronds of numerous Monogr apt idee, for no geologist, as 
far as I am aware, has described Monograptidcc with com- 
pound fronds, unless Dairon's specimens be taken as such. 
The early writers on graptolites looked upon the num- 
ber of stipes possessed by graptolites as a character of 
prime importance in defining genera, such forms as Dicho- 
graptus, Tetr agraphia, Didymograptus and Monograptus 


being largely characterised by the possession of eight, four, 
two stipes and one stipe respectively. In a recent paper 
by Professor Nicholson and the present writer (15) we 
have endeavoured to show that this is not the case, but 
that the character of the hydrothecae and to a less degree 
the amount of angle of divergence of the stipes are of im- 
portance. We endeavour to prove that certain grapto- 
lites underwent development along parallel lines, passing 
through many-branched, eight-branched, four-branched, 
two-branched and one-branched forms, thus illustrating the 
principle of heterogenetic homoeomorphy advocated by 
Mojsisovics, S. S. Buckman and others. If this be allowed, 
many of the present genera will have to be abolished and 
new ones formed ; but the writers earnestly advocate the 
retention of the present genera under existing circum- 
stances, and hope that the formation of fresh genera will 
be deferred until our views are more fully developed or 
perchance disproved, though we do not think that the latter 
event is likely. 

It will be noticed that the above researches into the 
morphology of the graptolites deal mainly with the 
celluliferous portions of the polyparies, whilst the study of 
the various bodies referred to as concerned in reproduction 
has not been largely pursued of recent years. 

Passing now to the memoirs treating of the graptolites 
as indices of age of the rocks which contain them, it may 
be remarked at the outset that recent work has fully estab- 
lished the correctness of the views advanced by Lapworth 
in his papers on the Moffat series and on the geological 
distribution of the Rhabdophora. Perner alone has stood 
out for the anomalous occurrences described by the eminent 
Barrande in the Bohemian basin, but he does not yet 
appear to have studied completely the zonal distribution of 
these organisms in that region, though he has added largely 
to the number of species occurring in the Lower Palaezoic 
rocks of Bohemia. The new species described here and 
elsewhere of recent years it is not contemplated to notice in 
this article, though they will doubtless give us much 
information in addition to that we have already obtained 


concerning the morphology and phylogeny of the graptoli- 
toidea. It would serve no useful purpose to give details of 
the numerous papers which confirm the value of the grap- 
tolites for purposes of correlation of the strata. In Britain, 
Lapworth himself has described a number of graptolitic 
bands interstratified with deposits containing the remains of 
other organisms in Ayrshire (16). Much remains to be 
done in this respect, for in order to utilise to the utmost the 
value of these organisms as stratigraphical indices, it will be 
necessary to have a complete correlation of graptolitiferous 
strata of all ages, with those which contain these organisms 
rarely or not at all. For this purpose all graptolites should 
be carefully collected and preserved from out of those 
deposits in which they are not frequent, and are associated 
with other organisms. They should be looked for especi- 
ally in calcareous deposits, for as we have already seen, such 
specimens are particularly valuable as furnishing information 
concerning the morphology of these fossils. The southern 
uplands of Scotland have recently been re-examined by the 
geological surveyors, and it is scarcely necessary to state 
that they have fully confirmed Professor Lapworth's classifi- 
cation of the Lower Palaeozoic Rocks of this region. In 
England Professor Nicholson and the present writer have 
defined graptolitic zones in the Skiddaw Slates, Llandovery, 
Tarannon, Wenlock and Lower Ludlow Beds (17). Messrs. 
Lake and Groom have detected the Monograptus gregarius 
zone of the Birkhill shales and zones of Monograptus per- 
sonalis, M. Flemingii, M. colonius and M. leint wardinensis 
near Corwen and Llangollen (18), whilst in a paper which 
has hitherto only appeared in abstract, Miss Wood and 
Miss Elles have detected several zones of the Birkhill-Gala 
beds near Conway. On the Welsh borderland W. W. 
Watts has found one graptolitic zone of Wenlock and two 
of Lower Ludlow age on the Long Mountain (19). In 
addition to this, various other graptolitic zones have been 
detected in different parts of Great Britain, and the zones of 
the Moffat area have been traced into Ireland. On the 
European continent, Linnarsson, Brogger, Tornquist, 

Tullberg and others have detected numerous graptolite 



zones in Scandinavia, a full account of which appears in 
Tullberg's paper on the graptolites of Scania (20), one of 
the most valuable of recent contributions to the literature of 
the graptolites. Tornquist, Perner, Barrois and others 
have also identified various graptolitic zones in Thuringia, 
Bohemia and France. In North America the principal 
contribution is by our own countryman, Lapworth, who has 
identified a number of graptolite zones in Canada, which 
are identical with those detected in Europe (21). In 
Australia T. S. Hall is studying the well-known Areing 
graptolite fauna, and finds that the graptolites here also are 
limited to special zones (22). A number of other papers 
might be quoted to show the general recognition of the 
utility of graptolites for purposes of correlation of strata, 
but enough has been said to indicate the manner in which 
the work is progressing, and the vast amount which yet 
remains to be done in this connection. I cannot leave this 
part of the subject without uttering a warning note. More 
harm is done by a wrong determination than good by a 
correct one. The graptolites are by no means easy of 
identification by those who have not made them a special 
study, and it is particularly desirable that no determination 
should be recorded by tyros, unless it is absolutely certain, 
for when once a wrong name has crept into a list it is 
exceedingly difficult to remove it. I could give several 
instances of very serious mistakes of this kind which have 
been made, each of which will have to be corrected else- 
where, but it would be invidious to give names in a 
general article of this character. 

We may now pass on to consider the physical conditions 
under which the graptolite-bearing strata were deposited. 
There is very little doubt that they were formed in water of 
very different degrees of depth, for graptolites are found in 
arenaceous, argillaceous and calcareous strata. Thev have 
mainly been collected from deposits which there is every 
reason to suppose were formed in deep seas, because a much 
greater number of individuals occur in a given space under 
such conditions than when the deposits were formed rapidly. 
The writer has elsewhere given cases of graptolitic deposits 


a few feet in thickness, being represented by thousands of 
feet in adjoining regions, and one naturally discovers 
forms more easily in a few feet of strata than in several 
thousand feet where the process of search rather closely 
approximates to that for the proverbial needle in the hay- 
stack. The evidence which is being gathered shows more 
strongly than ever that the thin graptolite-bearing shales, 
which for the above reasons have come to be looked upon 
as the deposits for graptolites /di?'- excellence, were deposited 
slowly in waters some distance from continents, and pro- 
bably of considerable depth. The evidence for depth 
depends mainly on the nature of the associated organisms, 
which are frequently dwarfed, and either blind or with 
enormously developed eyes, whilst that for deposition at 
a distance from land is confirmed by the ever-increasing 
number of cases of association of graptolitic deposits with 
others which are composed almost exclusively of tests of 
radiolaria. The most striking" case of this has recently 
been detected by the geological surveyors amongst the 
rocks of the Southern uplands of Scotland (23). Messrs. 
Peach and Home have there discovered beds with Tetra- 
graptus of Middle Areing Age, separated from beds with 
characteristic Glenkiln (Upper Llandeilo) graptolites by a 
thin deposit of radiolarian chert. " We thus perceive that 
the great mass of strata which elsewhere forms the Upper 
Areing, and the Lower and Middle Llandeilo formations 
are here reduced to not more than sixty or seventy feet. 
Judged by the palaeontological evidence these thin cherts 
appear to be a chronological equivalent of thousands of feet 
of ordinary sediment in North Wales. They, no doubt, 
were deposited with extreme slowness in a sea of some 
depth, and over a part of the sea-floor which lay practically 
outside the area of the transport and deposit of the terres- 
trial sediment of the time." 

The graptolites are generally viewed as type-fossils of 
the Lower Palaeozoic rocks, and this view is practically 
correct. The earliest graptolite which has hitherto been 
described, Dichograptus ? tenellns Linnrs., occurs in the 
Lingula Flags of Sweden, below the shales with Dictyo- 


graptus flabelliformis Eichw. which are so widely distributed. 
This Dictyograptus, by the way. which has a very limited 
vertical distribution, is probably in no way related to the 
long-ranged Dictyonema. Graptolites are extremely rare 
in the Upper Ludlow rocks, and have been detected in the 
Lower Devonian rocks of Bohemia, though it is doubtful 
whether their asserted occurrences in rocks of Devonian 
age in Scotland and the Harz Mountains are correct. It 
may be taken as fairly certain that they finally died out in 
Devonian times. Between the earliest and latest graptolitic 
deposits we have already a large number of graptolitic zones, 
which it will be of use to print in one connected list as this 
has not been heretofore done. So far as they have been 
made out they are, in ascending order, as follows : Lingula 
Flags ; (i.) Zone of Dichograptus? tenellus, Zone of Dictyo- 
graptus flabelliformis. Tremadoc Slates; Zones of Bryo- 
graptus. Areing Beds ; Zones of (i.) Dichograptus, (ii.) 
Tetragraptus, (iii.) Didymograpttts indentus var nanus. 
Llandeilo Beds; (i.) Zone of Didymograpttts Murchisoni, 
(ii.) Zone of Ccenograpttis gracilis. Bala Beds; Zones of 
(i.) Climacograptus Wilsoni, (ii.) Dicranograptus Clingani, 
(iii.) Pleurograptus linearis, (iv.) Dicellograptus complana- 
tus, (v.) Dicellograptus anceps. Llandovery Beds ; Zones of 
(i.) Diplograptus acuminatus, (ii.) Diplograptus vesiculosus, 
(iii.) Monograptus argenteus, (iv.) Monograptus convolutus, 
(v.) Cephalograptus cometa, (vi.) Monograptus spinigerus, 
(vii.) Rastrites maximus. Tarannon Beds ; Zones of (i.) 
Monographts turriculatus, (ii.) Monograptus exiguus, (iii.) 
Cyrtograptus Graycz. Wenlock Beds; Zones of various 
species of Cyrtograptus not yet fully worked out. Lower 
Ludlow Beds; Zones of (i.) Monograptus bohemicus, (ii.) 
Monograptus Alilssoni, (iii.) Monograptus leintzvardinensis. 
Upper Ludlozv and Lower Devonian ; Zones of undescribed 

It is quite certain that this number will be very largely 
increased as a result of further work, but it is sufficient to 
show the importance of the Lower Palaeozoic rocks when it 
is remembered that many of these Zones contain a fauna 
largely distinct from the faunas of the adjoining ones. 


When the Zones are worked out more fully than is the 
case at present, we shall have a far better gauge of " Geo- 
logical Time " than that founded upon the crude estimates 
made by measuring thicknesses of strata. 

Lastly, the study of graptolites may possibly throw 
some light upon climatic change. I have already en- 
larged upon this elsewhere (24), and pointed out that the 
separation of graptolitic deposits from non-graptolitic ones 
amongst the Stockdale shales of the Lake District, the 
deposits themselves being lithologically similar, is most 
readily explicable by climatic change. The argument 
would be stronger had microscopic examination and 
chemical analyses of the strata been made, and I should 
be glad to supply any one who cares to look into this 
question, which is one of some interest, with material for 
such examinations. 

In conclusion, the above notes will be sufficient to 
show the importance which the graptolitoidea have 
assumed not only to the geologist but also to the biologist. 
That they differ in any remarkable respect, as regards 
their teachings, from any other group of fossils is doubtful. 
Their special utility lies in the fact that owing to their 
characters they are preserved in sufficient numbers to 
allow collectors to obtain a large suite of specimens of 
almost every species with little difficulty ; the result is that 
further advance has been made in their study than in that 
of many other groups which like them are only preserved 
in the fossil state. One word to the biologists. We are 
often told that fossils are of little use on account of the 
absence of soft parts, though biologists have not been 
much hampered by this when dealing with the Vertebrata. 
But to compensate for the want of soft parts, we are furnished 
with a countless supply of specimens whose order of appear- 
ance and disappearance we are able to a large extent to ascer- 
tain, and this is what the biologist can never obtain by con- 
fining his attention to recent organisms. From them he has 
been able to ascertain that evolution occurs; how it occurs 
is left for the palaeontologist to describe. That the study 
of these organisms as pursued up to the present has not 


been in vain, is conclusively proved by the best of all tests, 
namely, that we are able to predict the discovery of forms 
which are afterwards detected by the worker in the field, 
to whom we commend this group as one specially worthy of 
his attention. 


(i) LAPWORTH, Charles. Notes on the British Graptolites and 
their Allies. I. On an improved Classification of the 
Rhabdophora. Geol. Mag., vol. x., pp. 500 and 555, 


(2) LAPWORTH, CHARLES. The Moffat Series. Quart. Journ. 

Geol. Soc, vol. xxxiv., p. 240, 1878. 

(3) LAPWORTH, CHARLES. On the Geological Distribution of 

Rhabdophora. Ann. and Mag. Nat. Hist., ser. 5, vol. iii., 

(4) Herrmann, Otto. Die Graptolithen familie Dichograptidae, 

Lapvv. Kristiania, 1885. 

(5) RlCHTER, H. Thiiringische Graptolithen. Zcit. d. Deutsch. 

Geol. Gesell., vol. v., p. 439, 1853. 

(6) SOLLAS, W. J. On the Minute Structure of the Skeleton of 

Monograptus priodon. Rep. Brit. Assoc., 1893, P- 7% l > 

(7) TORNQUIST, S. L. Studier ofver Retiolites. Aftr. nr Geol. 

Foren. i. Stockholm Forhdndl, Bd. v., 7, p. 292, 1880. 

(8) Perner, J. Etudes sur les Graptolites de Bohbne. Prague, 


(9) TORNQUIST, S. L. Observations on the Structure of some 

Diprionidae. Sdrtryck of Konl., Fysiogr., Svesk., Handl. 
Ny Folgd., 1892-3, Bd. iv. Lund, 1893. 

(10) Wiman, Carl. Ueber Diplograptidae Lapw., and Ueber 

Monograptus Geinitz. Bull. Geol. Inst., Univ. Upsala, vol. 
i., 1893. 

(11) Holm, G. Om Didymograptus, Tetragraptus och Phyllo- 

graptus. Aftr. ur Geol. Foren. i. Stockholm Forhdndl., 
1895, translated by Miss Wood and Miss Elles in Geol. 
Mag., vol. ii., pp. 433 and 481, 4th Dec. 

(12) Hall, James. Graptolites of the Quebec Group, 1865. 

(13) Dairon, James. Notes on Graptolites. Trans. Geol. Soc, 

Glasgow, p. 176, 1882. 

(14) Ruedemann, R. Synopsis of the Mode of Growth and 

Development of the Graptolitic genus Diplograptus. Amer. 
Journ. Sci., vol. xlix., 3rd ser., p. 453, 1895. 


(15) NICHOLSON, H. A., and Marr, J. E. Notes on the Phylogeny 

of the Graptolites. Geo/. Mag., 4th Decade, vol. ii., p. 529. 

(16) Lapworth, C. The Girvan Succession. Quart. Joum. Geo/. 

Soc, vol. xxxviii., p. 537. 

(17) Marr and NICHOLSON. On the Stockdale Shales. Quart. 

Joum. Geo/. Soc., vol. xliv., p. 654. Also Marr. On the 
Wenlock and Ludlow Strata of the Lake District, Geo/. Mag., 
3rd Dec, vol. ix., p. 534, and Notes on the Skiddaw Slates, 
ibid., 4th Dec, vol. i., p. 122. 

(18) Lake and GROOM. On the Llandovery and Associated 

Rocks of the Neighbourhood of Corwen. Quart. Joum. 
Geo/. Soc., vol. xlix., p. 426. And P. Lake, On the Denbigh- 
shire Series of South Denbighshire, ibid., vol. ii., p. 9. 

(19) Watts, W. W. The Geology of the Long Mountain on the 

Welsh Borders. Rep. Brit. Assoc., 1890, p. 817, 1891. 

(20) TULLBERG, S. A. Skanes Graptoliter. Sver. Geo/. Undersokn., 
_ ser. C, Nos. 50 and 55. 

(21) Lapworth, C. Preliminary Report on some Graptolites, etc. 

Trans. Roy. Soc. Canada, p. 167, 1886. 

(22) HALL, T. S. The Geology of Castlemaine, etc. Trans. Roy. 

Soc. Victoria, p. 57, 1895? 

(23) Geikie, Sir A. Annual Report of the Geo/ogicai Survey, etc., 

for 1895, p. 27, 1896. 

(24) MARR, J. E. On Homotaxis. Proc. Cambridge Phi/. Soc, 

vol. vi., pt. ii., p. 74. 

J. E. Marr. 


PART VI. (b). 

IN my article (59) on the flora of the African Islands of 
the Indian Ocean, I dealt with the subject in consider- 
able detail, but beyond the vascular cryptogams I had very 
few data concerning the Isle of Bourbon. Since then 
Dr. Cordemoy has published a Flora of the island (60), 
which is a consolidation of all the materials he has been 
able to collect during the leisure of upwards of thirty years' 
residence in the island, though unfortunately without a full 
collation with the rich earlier collections in the Paris Her- 
barium of Commerson, Du Petit-Thouars, and other botanists. 
Moreover, he has not worked out the geography of the 
plants to the extent he might have done, so that it takes 
some time to find and extract the particulars of special 
interest to the geographer. Indigenous and naturalised 
plants are included in the same enumeration without any 
typographical distinctions ; and the summary is limited to 
a table showing the number of species of each natural 
order, including naturalised species. A rough calculation 
of the number of indigenous species of vascular plants, 
described or enumerated, gives a total of about 11 00, 
whereof 200 are ferns, and 172 are orchids.' This is nearly 
250 higher than Baker's estimate (61) of the vascular plants 
of Mauritius; but, although the islands are nearly of the same 
size, the mountains of Bourbon rise to altitudes of between 
9000 and 10,000 feet, or about 6000 feet above those of 
Mauritius ; thus giving an additional climatic zone to the 
former island. And an analysis of the components of the 
flora shows that Bourbon possesses a much larger temperate 
element. But it should be known that Cordemoy takes a 
narrower view of species than Baker, especially in ferns ; 
and some allowance would have to be made for this in com- 
paring the totals. Apart from this divergence, the flora of 
the two islands is essentially the same, several genera and 
many species being common to both and found nowhere 


else. The predominating natural orders of vascular plants 
occupy nearly the same positions numerically in both islands ; 
ferns being first and orchids second, and Leguminosae and 
Compositae relatively low down ; very different proportions 
from those obtaining in the Madagascar Mora, in which 
these four orders occupy reversed positions. Thus : Legu- 
minosae, Filices, and Compositae, followed by the Orchideae, 
which are represented by just half as many species as the 

The absence of a number of natural orders from Dr. 
Cordemoy's Flora that are represented in Mauritius may 
be accounted for partly by the fact that he did not work out 
the old collections made before the destruction of the virgin 
forests which formerly clothed the island. It is probable 
that many species have disappeared from both islands from 
the same cause. The following orders known to be, or as 
having been, represented in Mauritius are not included by 
Cordemoy : Xyridaceae, Scitamineae, Podostemaceae, Myo- 
porineae, Bignoniaceae, Lentibulariaceae, Gentianaceae, 
Rhizophoreae, Connaraceae, Simarubaceae, Ochnaceae, Bur- 
seraceae and Nymphaeaceae. The absence of several of the 
foregoing orders might be accounted for without calling in the 
theory of destruction, but it would lead too far to attempt the 
discussion of the matter here. Myoporum mauritianum is 
an instance of a plant, and an order that is no longer repre- 
sented, if it ever were ; for there may have been an error 
in locality. The only specimen at Kew is labelled as coming 
from one small patch at the east end of the island of 
Rodriguez, which is some 300 miles distant from Mauritius. 
Moreover the Seychelles and Rodriguez between them 
possess several natural orders which do not reach Bourbon 
or Mauritius, though they are represented in Madagascar. 
They are Nepenthaceae, Passifloraceae, Turneraceae, Diptero- 
carpeae (?), Ternstrcemiaceae and Dilleniaceae ; whereof the 
first and the fourth are essentially Asiatic, the second 
and third American, and the two last equally Asiatic 
and American. The parasitical Rafflesiaceae are perhaps 
the only natural order in Bourbon that is not repre- 
sented in Mauritius. Cordemoy records Hydnora afncana 


as common at St. Paul in Bourbon. It inhabits Eastern 
tropical and South Africa, though it is not known 
from Madagascar or any other of the African islands. Six 
or seven species of Hydnora have been described ; all in- 
habiting Africa from Abyssinia and Angola southward to 
Cape Colony. I have previously noted (62) the discovery of 
a member of this order (Cytinus Baroni) in Madagascar. 
Since writing that I have seen a third Mexican species. 

The intimate relationships of the floras of Bourbon and 
Mauritius may be gathered from the presence in the two 
islands, and restriction to these islands, of the following 
monotypic, mostly very distinct, genera: Cossignya and Dora- 
toxylon (Sapindacese), Grangeria (Rosaceae), Roussea (Saxi- 
fragaceae), Psiloxylon (Lythracese ?), Fernelia (Rubiacese), 
Heterochcenia (Catnpanulacese), Bryodes (Scrophularineae), 
Monimia (Monimiaceae) Dictyosperma (Palmae). To these 
may be added several other genera of the same geographical 
area, represented by more than one species ; in five instances 
out of six by three species : Fostidia (Myrtaceae), Pyrostria 
and A/y 07itma (Rub'iacedz), Faujasia (Composite), Hyophorbe 
and Acanthophcehix (Palmae). Twenty-five other character- 
istic genera are restricted to the African region, using that 
designation in the sense of including therein the islands 
under consideration, Madagascar, and continental Africa. 
Trochetia (Sterculiaceae) is remarkable among them as 
extending to St. Helena, where it is represented by two 
distinct species — or rather was, for one is quite extinct in 
a wild state. Psiadia (Composite) has the same range. 

Allusion has been made (63) to the phenomenal con- 
centration of endemic palms in the Seychelles, and it would 
be interesting to give the distribution and affinities of the 
palms of the whole of the East African Islands ; but I must- 
confine myself to the Bourbon species. The native species 
are five in number, namely : Latania Commersonii, Hyophorbe 
indica, Dictyosperma album, Acanthophcenix rubra and A. 
crinita. All these palms also inhabit Mauritius, and 
they are, so far as our present knowledge goes, confined to 
the island. All the genera are peculiar to this insular region 
if we take in Madagascar, and Dictyosperma and Acan- 


thophcenix to Mauritius and Bourbon. Latania belongs to 
the Borasseae and all the rest to the Arecinese. As stated 
before, there is no parallel to this in insular floras of other 
parts of the world. Polynesia, both the eastern and western, 
is relatively poor in palms, and the West Indian Islands 
possess few endemic species ; but, as explained a few pages 
back, Lord Howe Island possesses four endemic species 
of palms belonging to Australian and endemic genera. 

Coming down to species we find that Cordemoy de- 
scribes about 200 new ones, which, with those previously 
known as endemic, would make probably not less than 25 
per cent, of the vascular plants endemic. It is probable 
that this number — the number of new species — maybe subject 
to some reduction, especially in such groups as the ferns 
and grasses in which so many species have a wide range ; 
yet 25 per cent, of endemic species is possibly below 
rather than above the mark. Nineteen grasses are de- 
scribed as new. Considering, however, the general distribu- 
tion of grasses, and that only four species are regarded as 
endemic in Mauritius, there are good grounds for suspect- 
ing that many of the Bourbon species are not really new. 

Orchids, epiphytal and terrestrial combined, contribute 
no fewer than seventy new species ; and the total number 
of orchids thus exceeds the total indigenous species of any 
other two natural orders. In Mauritius, orchids are more 
numerous than any other order of flowering plants, but 
they only occupy the first place by a majority of about ten. 
As I have shown elsewhere (64) orchids are exceedingly 
rare or entirely wanting in oceanic islands, and such pro- 
portions as Cordemoy's enumeration gives would hardly 
be found in the richest orchid districts of Asia or America. 
Continental tropical Africa, so far as known, is relatively 
poor, whilst in Madagascar, according to Baron's tabulation 
(65), orchids stand third, being exceeded by Compositae and 
Euphorbiaceae. It is true that I have estimated (66) that 
orchids are numerically more strongly represented in British 
India than any other order of flowering plants, and my 
estimate has proved correct in the subsequent elaboration 
of this order (67) by Sir Joseph Hooker. It may be in- 


teresting to add that orchids stand third in the flora of the 
whole world, and they also take the same position in the 
flora of Mexico and Central America. 

Returning to the Bourbon orchids ; the regional char- 
acteristic AngrcBCum is credited with eighteen new species, 
and a total of thirty-two species. There are also new 
species of the epiphytal genera Bulbophyllum, Aeranthus, 
and Saccolabinm ; but the bulk of the new ones are terres- 
trial plants, many of them very rare and inconspicuous, and 
most of them of short duration above ground. 

On this point Cordemoy says : " J 'en ai moi meme 
plusieurs nouvelles, en herbier, que leur mauvais etat de 
conservation ne permet pas de decrire. Certainement il 
en existe d'autres non encore decouvertes, surtout parmi 
les Ophrydees, dont plusieurs parcourent, en quelques 
semaines, la periode active de leur vegetation, puis se 
replient immediatement, pour passer le reste de l'annee sous 
terre a l'etat de tubercule. Plusieurs localites n'ont pas ete 
suffisament explorees.' r 

Three new genera of this group are described, namely, 
Acrostylia, Camilleugenia and Hcmiperis ; the first two 
being monotypic and the third having twenty-one species 
ascribed to it. All three would be included under Habenaria 
by some authors ; but in this extended sense Habenaria is 
a vast and heterogeneous agglomeration of species. 

Among other genera, of which several new species are 
described, I may mention Dombeya, Evodia, Eugenia, 
Embelia, Sideroxylon, Geniostoma, Psiadia and Faujasia. 

In addition to the new genera of orchids, four others 
are proposed, namely, Guya (Bixaceae), Herya (Celas- 
tracese), Allocalyx (Scrophulariacese), and Mahya (Labiatae). 
According to the author's own admission, three out of 
the four are somewhat doubtful, and the affinity of the 
fourth is not given more definitely than by placing it in the 
tribe Menthese. But Alahya stellata is an interesting plant, 
whatever its affinity, because it is believed to be the only 
really indigenous member of the Labiatae. It is a dwarf 
shrub, very rare, and found only near the summit of the 
Grand Benard, at an elevation of about 8650 feet. 


Strange to say the upper zone of vegetation is less 
alpine in character than that of the mountains of Mada- 
gascar and Tropical Africa. Cruciferae, Caryophylleae, 
Umbelliferae, Primulaceae and Gentianacese, as well as her- 
baceous Rosacese and Saxifragaceae, are either exceedingly 
rare or entirely absent. In Madagascar, where the highest 
point is barely 8500 feet, the following familiar genera 
occur : Ajuga, Alckemilla, Cattcalis, Crassula, D?-osera, 
Epilobium, Genista, Geranium, Linum, Pimpinella,Sanicula> 
Stachys, and Viola, besides many others which are unknown 
among the native plants of Bourbon. 

Gymnosperms are also unrepresented, both in the indi- 
genous vegetation, and among the numerous naturalised 
plants. It is the same in Mauritius ; but in Madagascar 
one species each of Cycas and Podocarpus has been dis- 
covered ; the latter being prominent in certain districts. 

Finally, I may add that the following orders are strongly 
represented in the Bourbon flora : Malvaceae (Dombeya, 2 1 
species, and Ruisia, Astiria, and Trocketia, regional 
genera) ; Rutaceae (Evodia) ; Urticaceae (Ficns, and Obetia 
and Maillardia, regional genera) ; Euphorbiaceae and Con- 

Since the untimely death of Dr. H. Baillon another 
part (68) of the admirable illustrations of the flora of 
Madagascar has appeared. It consists largely of plates 
for intercalation, and the highest number is 340. Unfor- 
tunately no descriptive or explanatory letter-press has been 
published in connection with these plates and none is likely 
to be forthcoming. Surgeon-Major H. H. Johnston has 
published (69) an enumeration of plants collected by him- 
self and regarded by him as indigenous in Mauritius, 
though they are not included in Baker's Flora. The 
total is fifty species, half of which are cellular crypto- 
gams. There is nothing specially remarkable amongst 
them. The same gentleman has published an account (70) 
of the vegetation of the small islands in the Mahebourg 
Bay, Mauritius, namely : He de la Passe, He Vakois, He 
aux Fouquets, He aux Fous, He Marianne and Rocher des 
Oiseaux. These islands are of coralline limestone forma- 


tion, and their flora is equally as poor, and composed mainly 
of the same species as the flora of the small coral islands of 
the Pacific Ocean, a specimen of which is given some pages 


(i) Science Progress, i., pp. 26-35. 

(2) J. T. Arundel. The Phoenix Group and other Islands of the 

Pacific, 8vo, pp. 8. Reprint from The New Zealand 
Herald, 5th and 12th July, 1890. 

(3) Report of the Eclipse Expedition to Caroline Island, May, 

1883. Memoirs of the National [American] Academy of 
Sciences, ii., pp. 1-146, with a number of views, etc. Botany, 
pp. 88-90, by W. Trelease, 1884. 

(4) C. M. WOODFORD. The Gilbert Islands. Geographical 

Journal, vi., pp. 325-350, with a map. Botany, p. 346, 1895. 

(5) W. Botting HEMSLEY. The Flora of the Tonga or Friendly 

Islands, with Descriptions of and Notes on some New or 
Remarkable Plants, partly from the Solomon Islands. 
Journal of the Linnean Society, xxx., pp. 158-217, tt. 9-1 1, 

(6) J. J. Lister. The Geology of the Tonga Islands. Quarterly 

Journal of the Geological Society, xlvii., pp. 590-617, with 
maps and views, 1891. 

(7) W. Botting Hemsley. Flora of the Solomon Islands. 

Kew Bulletin of Miscellaneous Information, 1 894, pp. 211- 
215; 1895, pp. 132-139; 1896, p. 17; and Journ. Linn. Soc, 
xxx., pp. 163-165, and 211-217, plates 9-1 1, 1894. 

(8) HOOKER'S Icones Plautarum, t. 17 14. 

(9) K. SCHUMANN. Die Flora von Kaiser Wilhelms Land, p. 69, 


(10) J. G. BOERLAGE. Handleiding tot de Kennis der Flora van 

Nederlandsch Indie, i., p. 445 et p. 673, 1890. 

(11) C. Hedley. The Range of Placostylus. Proceedings of the 

Linnean Society of New South Wales, ser. 2, vii., pp. 335- 
339, 1892. 

(12) Science Progress, i., p. 40, 1894. 

(13) H. N. Ridley. A Day at Christmas Island. Journal of the 

Straits Branch of the Royal Asiatic Society, No. 23, pp. 123- 
140, 1891. 

(14) H. Trimen. Handbook of the Flora of Ceylon, part ii., Con- 

naraceae to Rubiacese, 1894; part iii., Valerianaceae to 
Balanophoraceae, 1895. 


(15) H. Trimen. A Preliminary List of Maldive Plants. Journal 

of Botany, pp. 3-6, 1896. 

(16) Capt. VV. F. W. Owen. Geography of the Maldiva Islands. 

Journ. Roy. Geogr. Society, ii., pp. 81-92, 1832. 

(17) Science Progress, i., pp. 3 8 7-396. 

(18) T. Kirk. Trans, and Proc. N. Zeal. Inst, xxvii., pp. 327- 

359, i895- 

(19) W. Colenso, H. C. Field, and D. Petrie. Trans, and 

Proc. N. Zeal. Inst., xxvii., 1895. 

(20) A. Hamilton. Notes on a Visit to Macquarie Island. 

Transactions and Proceedings of the New Zealand Institute, 
xxvii., pp. 559-579, 1895. 

(21) Science Progress, L, p. 395. 

(22) Flora of Macquarie Island. Keiv Bulletin, p. 401, 1894. 

(23) T. KlRK. On the Flora of Macquarie Island. Report of the 

Third Meeting of the Australasian Association for the 
Advancement of Science, pp. 226-227, 1891. 

(24) T. KlRK. Description of New Grasses from Macquarie Island. 

Trans. N. Zeal. Inst., xxvii., p. 353. 
(25) 1 A. Vollmer. Lord Howe Insel, Pitcairn und Norfolk 
Insel. Petermann's Mittheilungen, xli., pp. 72--JJ, 1895. 

(26) C. MOORE. Sketch of the Vegetation of Lord Howe Island. 

A Report to the Under Secretary for Lands, New South 
Wales, 1869. 

(27) G. BENTHAM. Flora Australiensis. 

F. v. MUELLER. Fragmenta Phytograpihce Australia. 

(28) J. B. WILSON. Report to the Colonial Secretary on the Present 

State and Future Prospects of Lord Howe Island, 1 vol. 4to. 
With Photographs. List of timbers, p. 22. Vegetation by 
J. Duff, pp. 28-36, 1882. 

(29) ETHERIDGE. Lord Howe Island ; its Zoology, etc., 1889. 

(30) F. VON Mueller. Index omnium Insulae Howeanae Plantar- 

urn, quas hactenus obtenui, exclusis speciebus certe introduc- 
tis. Fragmenta Phytographia Australia, ix., pp. 76-78, 1875. 

1 Since my remarks on the flora of Lord Howe Island were put into 
type, I have seen the " Macleay Memorial Volume," published by the 
Linnean Society of New South Wales, in 1893. It contains a paper by 
Prof. R. Tate " On the Geographic Relations of the Floras of Norfolk and 
Lord Howe Islands," in which the author arrives at much the same conclu- 
sions as myself. I had also overlooked a short article by Mr. C. Moore 
{Transactions of the Royal Society of New South Wales, v. (1872), pp. 29- 
34), on the same subject, but dealing only with the distribution of the 
genera. I may add that my detailed account of the vegetation and flora 
of the island has appeared in the Annals of Botany for June, 1896. 


(31) O. Beccari. Malesia, i\, p. 66, 1877. 

(32) Science Progress, i., p. 400. 

(33) B. L. Robinson and J. M. Greenman. On the Flora of the 

Galapagos Islands. American Journal of Science, 1., pp. 
135-149, 1895. 

(34) W. BOTTING Hemsley. The Flora of the Galapagos Islands. 

Nature, lii., p. 623, 1895. 

(35) G. Baur. The Galapagos Islands. Proceedings of the Ameri- 

can Antiquarian Society, 1891 , and Reprint, 1892 ; and a fuller 
account in Bio/ogisches Centralblatt, xii., pp. 221-250. 

(36) G. Baur. The Differentiation of Species on the Galapagos 

Islands and the Origin of the Group. Biological Lectures 
delivered at the Marine Biological Laboratory of Wood's Ho 11, 
1894, pp. 67 -7S. 
($7) A. Agassiz. General Sketch of the Expedition of the 
Albatross. Bulletin of the Museum of Comparative Zoology 
at Harvard College, xxiii., pp. 1-89, plates 1-22, including 
two maps, 1892. 

(38) Botany of the "Challenger" Expedition, \., 1, Introduction, p. 5. 

(39) J. D. HOOKER. On the Vegetation of the Galapagos. Trans- 

actions of the Linnean Society, xx., pp. 235-262, 1847. 

(40) C. Darwin. A Naturalists Voyage, p. 393, edition of 1884. 

(41) A Naturalist's Voyage, p. 396. 

(42) W. BOTTING Hemsley. Cactacese in the Galapagos Islands. 

Nature, liii., p. 31, 1895. 

(43) Magazine of Zool. and Bot., i., p. 467, pi. 14, f. 2, 1837. 

(44) A Naturalists Voyage, pp. 374 and 382, edition of 1884. 

(45) Om Galapagos Oarnes Vegetation. Freg. " Eugenics" Resa. 

Bot., p. 95. 

(46) Biologischcs Centralblatt, xi., 1892. 

(47) J. D. Hooker. Botany of Ellesmere Land and Grinnell 


Land. Nares's Narrative of a Voyage to the Polar Sea, ii., 
pp. 301-310, 1878. 

(48) D. Oliver. List of Flowering Plants in Ellesmere Land and 

Grinnell Land, 8o°-83° N. lat. Nares's Narrative of a Voyage 
to the Polar Sea during 1875-6, ii., pp. 310-312, 1878. 

(49) Science Progress, ii., p. 379, 1895. 

(50) A. Trevor-Battye. Icebound on Kolguev. Notes on the 

Flora, pp. 396-417, 1895. 

(51) H. W. Feilden. Mild Arctic Climates. The Glacialists 

Magazine, [., pp. 91-95, 1893. 

(52) H. W. Feilden and H. D. Geldart. Notes on a small 

Collection of Spitzbergen Plants. Trans. Norf and Norw. 
Nat. Soc, vi., pp. 47-53, 1894. 


(53) F. J. Ruprecht. Flores Samoyedorum Cisuralensium. 

Beitrdge zur Pflanzenkunde des Russischen Reiches, zweite 
Lieferung, pp. 1-67, tt. 1-6, 1845. 

(54) H. MOHN. Einsamkeit Insel. Petermann's Mittheilnngen, 

pp. 57-59, 1879. 

(55) F. KURTZ. Verzeichniss der auf Island und den Faer Oern 

von Dr. Konrad Keilhack gesammelten Pflanzen. Abhand- 
lungen des Botaniscken Vereins der Piovinz Brandenburg, 
xxxvi., pp. 150-158, 1895. 

(56) J. D. Hooker. On the Canarian Flora as compared with the 

Maroccan. Hooker and Ball's Journal of a Tour in 
Marocco, pp. 404-421, 1878. 

(57) H. Christ. La Flore Ancienne Africaine. Archives 

des Sciences Physiques et Naturelles, xxviii., pp. 369-374, 

(58) I. Urban. Additamenta ad cognitionem Florae Indiae 

Occidentalis. Particula 2. Engler's Botanische Jahrbiicher, 
xix., pp. 562-681, 1894-5. 

(59) Science Progress, hi., pp. 447-75. 

(60) E. Jacob de CORDEMOY. Flore de file de la Reunion, 8vo, 

pp. xxvii. and 574, 1895. 

(61) J. G. Baker. Flora of Mauritius and the Seychelles, p. 15.* 

(62) Science Progress, Hi., p. 457. 
(6^) Science Progress, hi., p. 459. 

(64) Botany of the Voyage of the " Challenger," i., part i., Introduc- 

tion, p. 27, 1885. 

(65) Journal of the Linnean Society, xxv., p. 252, 1889. 

(66) Biologia Centrali- Americana. Botany, i., p. xv., 1888. 

(67) Flora of British India, v., pp. 667-910; vi., pp. 1-198, 1888-90. 

(68) H. Baillon. Histoire Naturelle des Plantes de Madagascar. 

Atlas, 36 e Fascicule, 1895. 

(69) H. H. JOHNSTON. Additions to the Flora of Mauritius. 

Transactions of the Botanical Society of Edinburgh, 1895. 
Reprinted with Additions, pp. 28, 1896. 

(70) H. H. JOHNSTON. Report on the Flora of the Outlying 

Islands of Mahebourg Bay, Mauritius. Trans. Bot. Soc. 
Ediu., 1895. Reprinted with Additions, pp. 36, 1896. 




QUESTIONS respecting the origin and development 
of race-types have been among the favourite 
battle-grounds of anthropologists since anthropology began 
to be. Some have held that the countless varieties of type 
in man could be accounted for by the simple admixture of 
a very few original types, of three for example, a white, a 
black and a yellow one, others that nothing was needed to 
produce the widest extremes of variation save the direct 
influence of what the French call "media" and the 
Americans environment. With the development and in- 
creasing prevalence of evolutionary theories, the questions 
were looked upon from a somewhat different point of view. 
The same two parties, however, continued to exist, the one 
assigning supreme importance to innate variability controlled 
by natural selection, the other to the same variability 
controlled by environment. In process of time it became 
obvious that there might be other selective agencies than 
those commonly understood by the term natural ; and Alfred 
Wallace himself pointed out that natural selection must 
have been potent in its working on man in the early stages 
of civilisation, but that in later stages it ceased to be so, 
while other agencies came into play. 

Questions dependent on, or arising out of those already 

mentioned are innumerable, and in some instances at least 

are of obvious and immediate practical importance. For 

example : Which are the types of man that are most 

suitable for colonisation or acclimatisation in different parts 

of the world ? and are they recognisable by colour or 

form of head, by kephalic or nasal index, by stature or any 

other visible character? What is the connection or relation, 

if any, between complexion and liability to malarial fever, 

to syphilis, to cancer or leprosy? Are the more fertile 

types or strains of mankind to be known by outward signs ? 

Are new types of man likely to be developed more suitable 

than those now prevailing to the altering conditions of civic 

and industrial life, and if so, through what agencies ? 


Let us begin with the subject of complexion or colour, 
because it is one of the most conspicuous differential char- 
acteristics of man. The xanthochroic type of Huxley, the 
blond, at present so dominant and aggressive, occupying, 
in conjunction it is true with the melanochroic (or dark 
white), more and more of those parts of the earth, such as 
North America, Australia and South Africa, which have 
hitherto been the patrimony of the brown or the black man 
— is there reason to expect that it will hold its own outside 
of its original habitat, or even there ? 

The historical evidence is on the surface at least un- 
favourable. If we take the words used in their most natural 
sense, we must allow that the Greeks and Romans de- 
scribed not only the Germans but the Gauls and Thracians 
as blond. And they did not mean simply that the blond 
complexion was pretty common among these northern 
people ; that could hardly have struck them as very re- 
markable ; for if they had not had among themselves pretty 
frequent examples of it, their descriptions of the four tem- 
peraments could hardly be explained. 1 Literary portraits, 
and personal names such as Flavius, Rufus, Ahenobarbus, 
leave no doubt that there was considerable variety of com- 
plexion among the Romans of the republican period, though 
dark hues may have prevailed ; and it does not appear that 
the continual influx of northern blood has been able to do 
much, if any, more than to maintain the status in that 
respect. The Greeks ascribed yellow locks to Achilles and 
Menelaus and other chieftains of the heroic age ; but in the 
imperial age the Egyptian limners represented Greek 
ladies with black hair and eyes. The ballads of Mount 
Rhodope, believed to be of extreme antiquity, and refer- 
ring to Philip, Alexander, and even Orpheus, ascribe yellow 
hair to their heroes ; but the Pomaks of the Rhodope are 
not now a blond race. 2 Another argument may be de- 
rived from the ancient Egyptian wall-paintings. Not only 

1 Among the marks of the sanguine and lymphatic temperaments light 
hair is generally mentioned, while black hair belonged to the choleric and 
the melancholic. 2 Fligier. 


the Lebo or Tahennu and the Amorites (both probably 
enough of North- European origin, though domiciled in 
Lybia and Canaan), but some of the Arabian Shashi are 
represented as of xanthous complexion. Yet now-a-days 
we hear nothing of blonds in the Arabian or Egyptian 
populations, except where recent admixture of blood may 
be suspected. Again, Flinders Petrie's recent discovery 
of the remains of a tall, brown-haired, and apparently 
"Aryan" population in Middle Egypt, 1 that seems to have 
completely and speedily disappeared, reminds us of the 
generally accepted statement that the Mamelukes have no 
representatives in the Egyptian population at the present 
day. On the other hand the alleged blond coloration of 
the Guanches in the Canary Islands, and the known fre- 
quency of that complexion in the people of the Riff, and in 
the Kabyles of some mountainous regions further east, 
make it probable that the type of the Tahennu still exists 
where climatic conditions are not unfavourable to it. And 
after all, these Tahennu may have been only a blond 
military aristocracy ruling a melanochroic plebs ; had it 
been otherwise, allowing that they had come from the 
north, why did they not perpetuate an Aryan language in 
North Africa ? 

Again, the large xanthous element in the Jews has been 
accounted for by the existence of an ancient Amorite cross ; 
and on the whole this appears to me the most probable 
explanation. We can hardly doubt its antiquity in any 
case, since it is present in every section of the Jewish 
people, and is very distinct among the Sephardim of the 
Levant, though perhaps larger in proportion among the 
Ashkenazim, 2 whose Gentile neighbours are so largely 
blond. In some parts of the Levant, indeed, among the 
dark Turks and Armenians, a red beard raises a suspicion 
of Hebrew ancestry. 

On the whole this kind of evidence, of which much 

1 " Indications of the earliest English occupation of Egypt," as 
De Lapouge pointedly remarked. 

2 Jacobs and Spielmann, Anthrop. Trans. Beddoe, Ethn. Trans. 


more might be adduced, leads me to think that though 
selective agencies in the warm Mediterranean regions are 
on the whole adverse to the perpetuation of the blond 
type, they are not so everywhere or in very high degree. 

Most of what evidence we have from northern countries 
makes one doubt whether any change has occurred except 
through immigration from melanochroic areas, and con- 
sequent admixture of blood. The Icelandic Sagas show 
that the Norsemen in the tenth century w T ere as diverse in 
colour of hair as they are now ; in fact the number of 
persons qualified as "black" would be a little surprising, if 
one did not allow for the probable inclusion of some whose 
hair was really only dark brown. The carefulness of the 
descriptions is vouched for by coincidences ; thus, chiefs 
with a mixture of Irish blood, such as Skarphedinn and 
Kjartan, betray it by some Irish feature. The eyes are 
seldom mentioned ; but Egil Skallagrimson, a pure Nor- 
wegian, had black eyes. 

Similarly the old Irish poems and legends testify to the 
occurrence of the same varieties of complexion that now 
exist, and particularly to that of the very Irish combination 
of blue eyes and black hair, which is ascribed among others 
to the famous Diarmaid O'Duibhne, the semi-mythical 
ancestor of the Campbells. 

Nevertheless I hold to the opinion, though only as an 
opinion, not as a firm belief, that the modern Norsemen 
are, if anything, more generally blond than their ancestors, 
and the modern Irishmen less so. If Scandinavia was, as 
now-a-days many think, the officina or breeding-ground of 
the blond long-headed type, may not the same agencies 
which worked in that direction after the close of the last 
glacial period be still operating there now, though it may 
be less powerfully ? As for the Irish, it is certainly curious 
that no early English writer, so far as I am aware, makes 
mention of their dark hair. As I have said elsewhere, 
Giraldus tells us that the Welsh were of swarthy com- 
plexion, but he says nothing about the colour of the Irish 
(though he had much to do with them), except that inci- 
dentally and casually he says something about " long 


yellow hair, like the Irish". The Irish colony about Dinas 
Mawddwy, in Merioneth, were called "the red men of 
Mawddwy". It is probable that the ruling tribes of 
Ireland had much more of the blond element than the 
servile ones ; 1 and that the former were exhausted by the 
long wars with the English, by the military emigrations to 
France and Spain, and perhaps the earlier emigrations to 
America. Dr. Morton, the first great American anthropo- 
logist, in describing the Irish as he saw them, said "eyes 
and hair light ". But there is no doubt that, speaking 
broadly, there is more dark hair in Ireland than in Eng- 
land or Scotland, though there are more dark eyes in 
England. The climate of Ireland, cloudy, moist and 
temperate, should favour the depigmentation of the eye by 
natural selection, and I have pointed out that the English 
colonists of Ireland by mixing their blood with that of the 
natives have changed their own type more in the direction 
of lightness of eye than of darkness of hair. 

Mr. Galton has pointed out how rapidly a community 
in which the age of marriage is late would, under like 
circumstances, be crowded out or superseded by one in 
which that age is some years earlier. This consideration 
is one of several which account for the rapid extinction of 
upper class families in these islands, while the proletariat 
multiplies with inconvenient rapidity ; and as the blond 
type is more prevalent in the upper than in the lower classes, 
it also is probably in process of diminution. If, however, 
it can be shown that the blond is more subject, in this 
country, to diseases of such a nature as to shorten life, and 
reduce the duration of the period of child-bearing and child- 
begetting, this same result would follow. Now there is a 
good deal of evidence as to the greater liability of blonds 
to certain classes of disease (in America at least), in 
Baxter's great work on the medical statistics of the Civil 
War. There are certain possible fallacies which may 
underlie Baxter's figures, to some of which De Candolle 

1 Thus MacFirbis, in a well-known passage, describes the Tuatha De 
Danaan as fair, and the Milesians as "white of skin, brown of hair," but 
the Firbolgs as a servile race, and black-haired. 


has directed attention ; but if we assume that the con- 
clusions which result from them are at all approximately 
Correct, it follows that the blonds in America have less 
chance than the brunets of contributing their due propor- 
tion to the next generation. Under these conditions the 
blonds ought to diminish relatively, and the brunets to 
increase ; and accordingly we find that of accepted soldiers 
there were among the white natives of the United States 
about (per cent.) 

66 light and 34 dark complexioned, but 
among the English 70 ,, 30 ,, 

Irish 70 „ 30 

„ Germans 69 ,, 29 „ 

Thus the men of American birth yielded a larger pro- 
portion of brunets than those of any of the nations that 
had most largely contributed to their ancestry, which is 
nearly equivalent to saying that the Americans are more 
generally dark complexioned than their ancestors were. 
Gould (quoted by Ripley) found that the natives of the 
eastern states were also darker than those of the west. 
But whether this last fact is occasioned by the parentage 
of the western men being more directly European, or 
whether it is connected with the more migratory character 
of the blond type, must be left for the present undeter- 

Of European evidence on the relation of complexion 
and disease there is, so far as I am aware, no great amount. 
My own observations have shown that it is a mistake to 
suppose, as many do, that light-haired persons are in 
England more liable to phthisis than others. I have also 
pointed out that cancer is more common in persons of dark 
complexion, and in this I am supported by the observations 
of Dr. Roger Williams. This last fact has, however, very 
little bearing on the subject in hand, for as cancerous disease 
usually attacks persons who are beyond the child-producing 
age it can have very little effect on the proportions of the 
different complexions of the next generation. 

As we possess for France not only elaborate recruiting 
statistics, with numerical lists for the principal disqualifying 


diseases, but also Topinard's departmental statistics of 
colour, and Collignon's of head-breadth, and Bertillon's 
of mortality, one ought, it would seem, to acquire therefrom 
some solid grounds for the connection of physical types with 
disease, and for the estimation of their comparative liability, 
and of the probable results in the direction of selective pro- 
pagation. In reality this turns out to be extremely difficult. 
"The prime difficulty" in such questions "is that these 
two factors, material prosperity and ethnic intermixture, 
in most cases follow the same laws of geographical dis- 
tribution." 1 

Thus in France the conquering races, in most of which 
blond types originally prevailed, occupied, as a rule, the 
most fertile tracts, which were also generally the most level 
and those contiguous to the great ways of communication. 
It is in such tracts that civilisation usually progresses fastest, 
that great cities arise with their vices and sanitary disad- 
vantages, and that blood is most mixed by continual 
migration and marriage. All these circumstances and 
conditions have to be taken into account before we can 
undertake to say anything as to the correlation of physical 
type with disease or military aptitude. The most promising 
plan seemed to me to be the throwing together of a number 
of departments having all one common character, but other- 
wise differing variously. The results thus gained are, 
however, more curious than conclusive. French anthro- 
pologists generally describe the tall, blond, long-headed 
type as subject to dental caries and myopia, and some add 
hernia to the list of its defects. Now the six departments, 
Nord, Pas-de-Calais, Somme, Aisne, Oise, and Calvados, 
which seem most distinctly to combine in their population 
all three marks of this type, have indeed a very bad record 
for dental caries, and, except Calvados, for general military 
unfitness ; but three out of the six stand much better than 
the average of France as regards myopia and hernia. More- 
over, bad teeth in the departments of France, strangely 
enough, usually co-exist with a low mortality, and I am 

1 Ripley, "Ethnic Influences in Vital Statistics," Q. P. American 
Statistical Association. 


disposed to think that both are the outcome of some 
influences which increase in potency with the advance of 
civilisation. In any case the frequency of dental caries 
does not seem to have an unfavourable selective influence. 

Phthisis, however, may and does have such an influence. 
And Houze, having shown that it is more prevalent among 
the taller and fairer Flemings than among the shorter and 
darker Walloons, concludes that it has been the principal 
agent in producing the supposed reduction of the blond 
type in Belgium and elsewhere. 

Now in England, as I have already stated, the propor- 
tion of blonds in the general population is quite as great as 
among the subjects of phthisis, but that of tall men among 
the phthisical is greater than that of short men. Let us see 
how it is in France. 

"Pulmonary disease," "scrofula" and weak "constitu- 
tion ' seem to be so often confounded or interchanged in 
the recruiting statistics, that I have thought it advisable to 
class the three together, with the following results. 

The three together are, or rather were in Boudin's 
time, the cause of rejection of conscripts in about the follow- 
ing order : — 1 

49 in France. 

42 ,, 10 most blond departments. 

39 ,, 10 most brunet. 

41 ,, 10 departments with tallest population. 

44 ,, 10 ,, ,, shortest ,, 

54 ,, 6 ,, ,, combination of stature, 

blond complexion and long head. 
48 ,, 5 departments with combination of stature, 

blond complexion and long head, 
excluding the Nord. 

1 This is not the correct way of putting it ; but we have here the result 
of averaging the ranks in each of the three classes of disqualification, and 
counting each of equal value. In reality the number rejected for weak- 
ness of constitution is vastly greater than that for scrofula, and that again 
than for phthisis. 

The low position of France as compared with her components is due 
to the greater and denser population of some of the vvorst departments, 
such as Seine and Nord. 

43 in 10 


, 10 


, 10 


• 5 


- 5 

22 , 

- 4 


. 9 


1 U 


in 10 departments most long headed. 

,, most mountainous. 

,, most level but thinly peopled. 

,, most urban. 

,, Normandy. 

,, Brittany. 

,, with population of Auvergnat type. 
„ ,, ,, ,, Remolothringian 

Unquestionably the northern blond type does show badly 
here, but whether the blond complexion is much in fault 
is doubtful. The Remolothringian region (= Austrasia, 
or Champagne and Lorraine), which is one of the most 
blond areas in France, but brachykephalic, stands extremely 
well ; in Brittany the Morbihan, the most blond depart- 
ment, stands best, and in Normandy the Orne, the least 
blond, stands worst. 

The low position of the Nord may be compared with 
that of the ethnologically similar or almost identical Flemish 
zone of Belgium. Houze himself ascribes this partly, but 
not, I think, wholly, to poverty, crowding, sedentary 
occupation, in fact to a number of causes outside of 

Another method of inquiry suggests itself. If it be 
true that the blond type is more susceptible than the 
brown to the malign influences of urban life, and especially 
to phthisis, which is largely a disease of crowded city- 
dwellers, we should find this type less frequent proportionally 
than the brown in ancient cities. On this point we have a 
great deal of evidence; the greater part of this is supplied by 
the great inquest of Virchow into the colours of the school- 
children of Germany, those of Schimmer in Austria, of 
Kollmann in Switzerland, and of Vanderkindere in Belgium : 
we have also the observations on adults in Italy of Livi, 
and those of myself in the British Isles. 

Georg Mayr, analysing the returns for Bavaria, 
pointed out that the town populations had on the whole a 
larger proportion of dark eyes, hair, and complexions than 


the rural districts, and it appeared to him that this excess 
could not be accounted for by the larger proportion of Jews 
in the towns, as it occurred, though perhaps to a less extent, 
in places where the Jews were few. 1 

The subject has not been so carefully worked out for other 
parts of Germany ; but a cursory examination of Virchow's 
figures shows that there is a larger proportion of dark hair 
in most of the great cities than in the surrounding rural 
districts, and this is more decidedly the case with the pro- 
portion of brown compared to blue eyes. Of 32 urban 
communities I find that in 

1 1 the proportion of dark hair to fair is greater, and that of brown 

eyes to blue much greater than in the surrounding districts. 
4 — of dark hair greater, of brown eyes greater. 
4 — of dark hair greater, of brown eyes greater in less degree. 
2 — of dark hair equal, of brown eyes greater. 
2 — of dark hair less, of brown eyes greater. 

4 — of dark hair about equal, of brown eyes about equal. 

5 — of dark hair less, of brown eyes less. 

These last are Halle, Wiesbaden, Krefeld, Ulm and 
Metz, most of which are towns which have grown rapidly 
of late. In the case of Metz the recent additions to the 
population have been derived from the blond region of 
Northern Germany. It may be noted that it is in that same 
blond region, generally speaking, that the most marked 
examples of the rule just laid down occur, 2 which fact 
strengthens the suspicion that the phenomena are largely 
due to the fact that the populations of these cities are 
partly constituted by immigrants of dark complexion from 
southern countries, including the Jews. 

In Schimmer's Austrian statistics this last source of 
difficulty is avoided, the Jews being returned and classified 
separately from the Gentiles. Of 30 cities separately re- 
turned, 1 5 show a larger percentage of dark hair than their 
surrounding districts, and 14 a smaller one; in the remain- 

1 Thus it does appear in non-Semitic Nurnburg, though it is much 
more distinct in Semitic Furth. 

2 E.g., Minister, Hanover, Altona, Berlin, Posen, Danzig, Elbing, 


ing one, Linz, the proportions are identical. So far, then, 
there is blank disappointment ; but when the eyes are 
examined the case is quite different : 27 cities show a larger 
proportion of dark eyes than their environs, and 3 only a 
less proportion. 1 In several of these 27 cases questions of 
race at once suggest themselves. In the Czechs, as in the 
Irish, the combination of light eyes with dark hair is 
common, while it is rare among the Germans. When, 
therefore, we find that in all the 6 cities of Moravia 
German is the school language, while in the country 
districts it is either Slavonic or mixed, and that in every 
one of these cities the eyes are darker and the hair lighter 
than in the surrounding districts, we need go no further for 
an explanation. But this will not serve in all the cases ; 
and some probability remains that there is a certain kind of 
selection at work to darken the eyes of the urban popula- 

In Belgium the case is not so clear. Ghent, Antwerp, 
Ostend, and Verviers come out much darker than their 
neighbours ; in most other cases the differences are slight 
either way. The cantons in Belgium are generally large, 
so that it is difficult to separate the urban and the rural popu- 
lations. I have however picked out 1 1 cantons in which I 
think the urban element most greatly preponderates, and the 
results are as follows. 

Of the 11,9 have a larger percentage of dark eyes than 
the arrondissements to which they belong, 1 of lighter eyes ; 
and in one, Mechlin, there is equality. But the hair, as in 
Austrian schedules, comes out about equal ; in 6 of the 
towns it is darker ; in 5, including Brussels and Liege, it is 
lighter. Ghent is, I suppose, the city in which the unfavour- 
able selective influences of urban life (overcrowding, poverty, 
sedentary occupation, infectious disease, etc.) are likely to 
have been most intense. 

In Switzerland Dr. Kollmann's schedules yield only two 
instances of a nearly pure urban community, Basel and 

1 Dozen, Bielitz, and Czernowitz (in the Bukowina) : they are all 
comparatively small places, and all near to race frontiers, which may 
possibly account for the anomaly. 


Geneva. Each of them is on a frontier, each is a singularly 
favourable specimen of a city, and is of little service for our 
purpose. Both Basel and Geneva have almost certainly a 
more blond population than that which surrounds them, 
whether Swiss, French, or German. 

In the West of England, according to my own published 
observations on 3630 adults, mostly hospital patients, of 
whom 2486 were natives of towns, and 1144 of rural 
districts, the proportion of dark hair in towns was to that 
in the country, reckoning by the index of nigrescence, as 
31 to 35 ; but that of dark eyes was as 58 to 49. We have 
here nearly the same phenomena as those we found to be 
so common in Germany, Austria, and Belgium. 

In the British Isles generally, the drift of my own very 
extensive local observations (in which the place of birth 
however was never actually ascertained) was to show that 
in large towns, especially those with an old settled popula- 
tion, the darker colours both of hair and eyes were more 
prevalent than in the surrounding districts. This applied 
to the greater part of Britain, but in parts of the west where 
the native population is generally dark-haired, e.g., Shrews- 
bury and Truro, the proportions may be reversed. The 
British military statistics, so far as investigated, viz., to the 
number of 13,800 deserters, yield results similar, but not 
strongly marked. Thus London, Birmingham, Bristol, 
Newcastle, Brighton, and Portsmouth give an index of 
nigrescence of 8, against one of 4*9 for the rest of England ; 
the proportion of dark eyes for the towns named being 39*5 
per cent., but for the rest of England, 387. Edinburgh 
and Glasgow give together an index of 1 1 '8, the rest of 
Scotland of 0*3 only, the percentages of dark eyes being 29 
and 27*8 ; and Belfast and Dublin give an index of 187 
against 15*2 in Ulster and Leinster, with percentages of 
dark eyes amounting to 32 and 28*4. The figures might 
be dissected with advantage, but to do so would lengthen 
this paper inordinately. 

Livi's statistics as to this point are perhaps the most 
interesting, and have the advantage of beino- founded on 
the physical characters of adolescents (i.e., conscripts). He 


finds that fair hair is more uncommon and dark eyes are 
more frequent among the inhabitants of cities and their 
immediate vicinity than among those of the surrounding- 
country. And this applies more or less to the whole of 
Italy, and cannot, therefore, apparently be accounted for 
by the immigration of the dark type from southern Italy 
into the northern cities, where the blond type is more 
common than in the south. 

Thus I find in the northern and more blond region 
(Piedmont, Lombardo-Venetia, Liguria) 17 urban popula- 
tions which, on a balance of eyes and hair, are darker than 
the rural populations around ; 3 which are lighter, Brescia, 
Como, Rovigo ; and r, Verona, where the conditions are 
equal. In the central provinces, from Emilia to Campania 
inclusive, 19 cities are darker, 9 are lighter, and 2 are equal. 
In the south, including Apulia, etc., and the islands, where 
blonds form a very small minority, 1 1 cities are darker and 
5 lighter. Thus in the north the rule obtains in 82 per 
cent, in the centre in 63, in the south in 69. The greater 
darkness appears to affect the eyes and the hair with 
something like equality, though not uniformly. 

Livi, finding that the blond complexion is, with identity 
or supposed identity of race, more prevalent in the poverty- 
stricken mountainous districts than in the plains, and putting 
that fact into connection with its less prevalence in the 
cities, is disposed to consider it as connected with poor 
food and hard labour, which may retard development of 
pigment ; in fact, he thinks the deposition of pigment to be 
an index of force and of development. Of course this is as 
yet unproven, and there is much to be said for and against 
the doctrine. But it does seem that we have evidence 
enough to show that in a great part of Europe the citizens 
are darker than the peasantry. This may be due to some 
direct influence of urban life, such as deficient oxygenation 
of the blood in children, but that seems very improbable. 
More probably it is due either to some kind of social 
selection such as Ammon and De Lapouge have studied, 
or else to the selection of the fittest for town life by the 
destructive agency of conditions more unfavourable to the 


blond than to the brunet child. I propose to follow out 
the subject further in another article. 


Baxter. Medical Statistics of the Provost- Marshal-General 's 

Bureau. Washington, 1875. 
BEDDOE. Races of Britain, 1885. Test, of Local Phen. in IV. of 

England to Permanence of Anth. Type. Anth. Me?n., vol. ii. 
HouzE". La Taille, etc. Bruxelles, 1888. 

KOLLMANN. Statist. Erheb. u. d. Farbe der Augen, etc. 1881. 
BOUDIN. Geographie Medicate. Paris, 1857. 
Livi. Antropometria Militare. Roma, 1896. 
SCHIMMER. Erheb. u. d. Farbe der Augen, etc. Oesterreichs, Wien, 

VANDERKINDERE. Nouvelle [Recherches sur V Ethnologie de la 

Belgique. Brux., 1879. 
VlRCHOW. Gesammtbericht . . . uber die Farbe, etc., der 

Schulkinder in Deutschland. Archiv f Anthr. 1886. 
GEORG Mayr. Die Bayerische fugend. 1875. 

John Beddoe. 


FOR reasons long ago pointed out by Lyell, fossil 
remains of birds are much more rarely found than 
those of other vertebrates, and, as a rule, occur in a very 
fragmentary condition. These circumstances, coupled with 
the difficulty in arriving at accurate determinations, owing 
to the great general similarity in the skeletal structures in 
most of the members of the class, have a direct bearing 
upon the scantiness of the results that have been attained 
in avian palaeontology. In spite of these drawbacks, how- 
ever, some not inconsiderable additions to our knowledge 
of fossil birds have been made during the last two or three 
years, and a short account of the chief papers on this subject 
may be of some interest. It will be convenient to take the 
papers roughly in the order of the geological age of the 
fossils they treat of, and to commence with those relating to 
the most ancient types. 

Pre-tertiary Birds. — Unfortunately, with one exception, 
no remains of pre-tertiary birds have been discovered during 
the last few years. This is the more to be regretted because, 
interesting though many of the tertiary birds may be, they 
are in all essential respects similar to recent forms, and 
throw no light whatever on the mystery of the origin and 
early history of the group, the key to which lies buried in 
the Jurassic and Cretaceous rocks. 

The single exception referred to is an imperfect tibia 
obtained at Judith River, Montana, from Cretaceous 
deposits of somewhat later date than those which formerly 
yielded the remains of Hesperornis and Ichthyornis. This 
tibia has been described by Marsh (i), who regards it as 
indicating a bird about two-thirds the size of Hesperornis, 
to which it is closely related, and has made it the type 
of a new genus, Coniornis, its specific name being C. alius. 

Tertiary Birds. — In tertiary deposits of various ages 
and in widely distant localities, some important discoveries 
of bird remains have been made of late. 


From the Eocene of New Jersey Marsh (2) has described 
some fragmentary bones which he considers belonged to a 
large struthious bird, Barornis, related to Gastornis and 
Diatryma from the Eocene of Europe and North America 
respectively. The specimens seem, however, to be too im- 
perfect to admit of complete certainty as to the affinities of 
this bird, but it may be remarked that the " struthious " 
nature of Gastornis is very doubtful, though it was pro- 
bably " ratite ' : in the morphological sense of that much 
abused term. 

A portion of a metatarsus obtained in Vancouver Island 
from a deposit of Eocene or, at latest, Oligocene age, forms 
the subject of a memoir by Cope (3). This author, after 
an exhaustive comparison with recent types, comes to the 
conclusion that its affinities lie in the direction of the 
Steganopodes, and that of these Pelecanus is the nearest 
ally of the extinct form, to which the name Cyphornis 
magnus has been given. The presence of a large pneumatic 
foramen on the anterior face of the bone is strongly in favour 
of this view, and if, like the Pelicans, Cyphornis was 
capable of flight, it is by far the largest flying bird hitherto 

A most important addition to our knowledge of the 
avi-fauna of the earlier tertiary rocks of Europe has recently 
been made by Professor Milne Edwards, to whom students 
of this branch of palaeontology are already more deeply 
indebted than to any other writer. In a paper (4) read at 
the Ornithological Congress at Buda Pesth, he described a 
number of bird remains from the well-known deposits of 
phosphate of lime (Phosphorites) which occur in the neigh- 
bourhood of Caylus (Lot) in Southern France. The mam- 
malian fauna of these deposits, described by Filhol and 
others, is an extremely rich one, and Lydekker has shown 
that several characteristic members of it occur at Hordwell 
in Hampshire in strata of Oligocene (Up. Eocene) age. 

The birds now described belong to some seventeen 
genera, of which ten are new ; these include representatives 
of several sub-orders. Only the more interesting of the 
new forms need be noticed here. 



Of these perhaps the most important is Archczotrogon, 
which is closely related to the Trogons, and may indeed be 
an ancestral form of the genus Trogon, an extinct species 
of which has been recorded from the Miocene of Allier. 
At the present day these birds occur in the Neotropical, 
Ethiopian, and Indian regions ; and it is remarkable that 
the extinct Miocene bird of Southern France should belong 
to a Neotropical genus rather than to one of those found in 
the Old World. This peculiar distribution of the recent 
and fossil forms is shown in a still more marked manner 
in the case of the next genus, Filholornis, which is said 
to be closely allied to Opisthocomns. The only known 
representative of this genus is the Hoatzin (Opisthocomus 
cristatus), which is one of the most peculiar and isolated 
forms of Carinate birds now living. It occurs only in 
Guiana and the Amazonian region, and is referred to a 
separate sub-order of which it is the only member. It is 
usually regarded as a primitive type, and the occurrence 
in Europe of a closely related bird is, therefore, another of 
those numerous cases in which such generalised types, now 
found only in the Southern hemisphere, have extinct re- 
presentatives in the Northern. That the determination of 
the affinities of Filholornis is correct there seems to be no 
doubt, since Milne Edwards states that its ulna is almost a 
facsimile of that of the Hoatzin, in which that bone is of a 
peculiar and distinctive form. 

In the same memoir several new ralline birds are added 
to the already numerous rails recorded from the Tertiaries 
of France. One of the new forms, Rallus dasypus, though 
much smaller, is said to resemble Ocydromus in the form of 
its humerus ; and another, Elaphrocnemus, a new generic 
type, approaches Aphanapteryx in the structure of its meta- 
tarsus. The occurrence in the lower Tertiary deposits of 
Europe of a large number of rails seems to be rather a 
strong argument in favour of a northern origin of the group, 
which, as Milne Edwards points out, is an extremely ancient 
one, of which at the present day we are only acquainted with 
some more or less degenerate descendants. Many of the 
more modified forms, such as Ocydromus and Aphanapteryx, 


are now confined to, or have recently become extinct in, the 
Southern hemisphere. Between these and the primitive 
generalised rails there must have been many intermediate 
forms, one of which, in the opinion of Milne Edwards, is to 
be found in this new genus, Eldphrocnemus. 

Other new genera of which the affinities are more doubt- 
ful are Orthocnemns, which resembles the Storks and Bus- 
tards in some respects and the Rails in others, and Tapinopus, 
which seems to have been a short-legged wading bird. We 
may also notice Necrobyas, a genus of owls presenting a 
combination of characters not found in any recent form ; 
TacJiyornis hirundo (previously described by Lydekker as 
Aigialornis gallicus), which is referred to the Cypselidce ; 
Dynamopterus velox, a cuckoo closely resembling Eudynamis 
orientalis, an inhabitant of the Austro-Malayan region ; 
Geranopterus, allied to the Rollers and Momots ; and, 
lastly, Pterocles vatidus, a sand-grouse considerably larger 
than any recent species. 

Although many of the genera and species above noticed 
are founded on single or, at best, a very few bones, still in 
the hands of one so experienced in avian osteology as 
Professor Milne Edwards such material is sufficient for 
a fairly certain determination of the affinities of the fossil 
forms ; and in this case the importance of the results from 
the point of view of geographical distribution cannot easily 
be over-estimated. It is much to be regretted that this 
valuable paper is not illustrated, since even the most careful 
descriptions of bird bones are very unsatisfactory without 

From the Middle Miocene of La Grive-St.-Alban in 
South-Eastern France, Lydekker (5) has described a 
small collection of bird bones. These, which do not in- 
clude any very striking novelties, are referred to a new 
species of Owl, a large Pheasant, previously recorded by 
Milne Edwards from beds of about the same age at 
Sansan, a number of quail-like birds {Palcsortyx\ a Sand- 
piper and an undetermined Picarian bird. 

The next addition to the ranks of fossil birds to be 
considered is by far the most important that has been 


made since Marsh's discovery of the Toothed birds of 
North America. In this case Patagonia, a region long 
known for the wealth and peculiar character of its fossil 
Mammalia, has yielded a number of the most extraordinary 
avian types yet known. The discovery of these is due 
to Dr. F. Ameghino and his brother, to the former of 
whom we are indebted for the most complete account of 
them that has yet been published. 

The first mention of the existence of gigantic extinct 
birds in Patagonia occurs in a letter from Carlos Ameghino, 
published in the Revista Argentina de Histoida Natural, 
April, 1 89 1, and containing a report of the results of his 
collecting expedition in Patagonia. 

Some years before this (in 1887) F. Ameghino (6) had 
described under the name Phororhacos longissimus the 
symphysial portion of a large mandible which he considered 
to belong to an edentate mammal ; a portion of a cranium, 
the type of the genus Tolmodzts, was also referred to 
a member of the same class. In 1891, however, thanks to 
the new and better material obtained by his brother, he 
was able to show clearly (7) that both these specimens were 
in fact portions of the skeletons of gigantic birds, and to 
give a fairly complete diagnosis of the genus Phororhacos. 
In some points, as for instance in the statement that teeth were 
present, and that there was a bony helmet-like crest on the 
skull, this diagnosis, as Ameghino himself afterwards showed, 
is not quite correct; but it was the first definite statement 
of the chief characters of these extinct birds. The mandible 
was shown to be of enormous size and to curve upwards 
at its anterior end in a manner almost unique among birds ; 
for though Psophia and Dicholophus were compared with 
it in this respect, they do not in fact possess this character. 
The upper mandible forms a strong hooked beak like that 
of a raptorial bird. 

In the same year ( 1S91) Moreno and Mercerat pub- 
lished a catalogue of the fossil bird remains in the La 
Plata Museum (8). This was illustrated by a large series 
of very beautiful photographic plates, but unfortunately 
these were unaccompanied by any adequate description of 


the specimens. Several extinct penguins of the genus 
Palceospheniscus, from so-called Oligocene beds, as well as 
a number of Pleistocene bird remains, were figured in this 
work, but by far the most important section is that dealing 
with the great flightless birds of the Santa Cruz Beds. 
For the reception of these the authors established a new 
order, the Stereornithes, which was subdivided into four 
families, the Brontomithidae (including the genera Bron- 
tornis and Rostromis), the Stereornithidce (with Phororhacos, 
Stereornis Mesembryornis and Patagomis), the Dryor- 
nitkidce (with Dryornis) and the Darwinoriiithidce (with 
Darwinornis and Oweniornis). Psilopterus (a name which, 
being preoccupied, was afterwards changed by Ameghino 
to Pelecyornis), a genus probably related to Phororhacos, 
was placed in the Cathartidce. 

At the end of the same year Ameghino published a 
synopsis of the South American fossil birds (9) in which 
he severely criticised the classification given above. He 
asserts that nearly all the new genera are merely synonyms 
of Phororhacos ; the only exceptions being Brontornis 
which includes Rostromis, and Psilopterus {Pelecyornis) 
which embraces Patapornis. Examination of the figures 
given by Moreno and Mercerat shows that in many 
cases, at least, he is right ; for instance Ste?'eomis is 
clearly the same as Phororhacos. On the other hand 
Dryornis, the sole member of the Dry omit hidce, is founded 
on the distal end of a humerus, which, judging from the 
figure, is probably that of a large vulture, most likely the 
Condor ; it may be pointed out that this specimen is not 
from the Santa Cruz Beds but from a Pleistocene de- 

In this paper Ameghino himself refers all these flightless 
birds to two families, the Pelecyornithidce (including Pele- 
cyornis and two new genera, LopJiiornis and Anissolomis) 
and the Phororhacosidce (with Brontornis, Phororhacos, and 
a new genus, Opisthodactylus). All these he regards as 
Ratitce, and in this he was followed by Gadow (10) and 
Lydekker (11). Subsequently the latter of these writers, 
relying on the fact that the quadrate in Phororhacos pos- 


sesses a double head for articulation with the skull, changed 
his opinion, and now considers them as degenerate Carinatse 
in which the wing has been reduced in size. 

Till recently our knowledge of the Stereornithes depended 
almost entirely on the preliminary notices of Ameghino, 
and on the plates of the catalogue of Moreno and Merce- 
rat. At the beginning of last year however the former 
author published by far the most important contribution 
(12) to this subject that has yet appeared. He now de- 
scribed not only the specimens to which his preliminary 
notices referred, but also a large number of additional 
remains. The classification followed in this paper is dif- 
ferent from that in his " Enumeracion," the order Stereor- 
nithes being adopted and subdivided into two families, the 
Phororhacidcz and Opisthodactylidce. In the former Pele- 
cyornis and Lophior'nis are now included, while Anissolornis 
is considered to be a Gallinaceous bird : several new genera, 
some of which appear to be of rather doubtful validity, 
are also added. The Opisthodactylidce include one genus 
only, Opisthodactyhis. 

Of the Phororhacidcz the skeleton of Pkororkacos injlatus 
is by far the most completely known, the skull, mandible, 
pelvis, the bones of the fore and hind limb, and some 
vertebrae being described and figured. The skull is of 
a very remarkable appearance ; from the side it most 
resembles that of a Raptorial bird, the enormous beak 
being sharply hooked at the anterior extremity, but when 
looked at from above it is seen to be much compressed, 
so that the premaxillary region, though very deep from 
above downwards, is extremely narrow from side to side. 
The quadrate has a double head for articulation with the 
skull, a character which, as Lydekker has pointed out (13), 
is opposed to the inclusion of these birds in the Ratitce. 
The mandible is very heavily built, and its anterior end is 
curved upwards in a manner very unlike the ordinary avian 
mandible. The sternum is, unfortunately, quite unknown, 
but the coracoid and scapula have been preserved. The 
former is long and slender, quite unlike that of any Ratite 
bird ; the acro-coracoid process is almost entirely wanting, and 


the only avian coracoid which at all resembles the fossil in the 
form of its upper end is, I believe, that of Aptornis. The 
wing-bones are very small in proportion to the size of the 
bird, but, at the same time, are stout and strong ; the ulna 
bears a number of tubercles marking the points of insertion 
of the secondaries. The pelvis is long and narrow, but in 
the posterior half, at any rate, it has been somewhat crushed, 
so that in fact it is broader than would appear from Ame- 
ghino's figure. The hind limb is long and comparatively 
slender ; in the tibia there was a bony extensor bridge, and 
in the metatarsus the hypotarsus is simple. All the above 
details are taken from the skeleton of a single individual of 
the smaller species, Phororhacos inflatus, in which the skull 
is about thirteen inches long. In these birds the head is 
proportionately very large, and this species probably only 
stood about three feet high at the middle of the back. 
Phororhacos longissimus is about twice as large, the skull 
being two feet long and about ten inches high. Of the 
other genera Pelecyomis is the best defined, the pelvis and 
most of the limb bones being known. As already mentioned 
this genus was placed by Moreno and Mercerat among the 
Cat hart idee : and though there is little doubt that this is 
incorrect, it is by no means clear that Ameghino is justified 
in placing it in the Phororhacidce, the pelvis being strikingly 
different from that of Phororhacos and the wing proportion- 
ately so much larger that it was probably still efficient as an 
organ of flight. The other genera of the family are for the 
most part known only from mere fragments of limb bones. 
Brontornis is a much larger and more heavily built bird than 
the largest species of Phororhacos, and Opisthodactylus is 
chiefly remarkable for the peculiar position in which Ame- 
ghino supposes the hind toe to have articulated with the tarso- 
metatarsus. In this paper also several extinct Penguins 
are described, as well as a number of ordinary Carinate 
birds belonging to several families. 

Lydekker points out (14) that the age of the deposits in 
which these avian remains are found is probably much over- 
estimated by the South American writers, and that they are 
probably Miocene. He also discusses the relationship be- 


tween the Gastornithidce of the Eocene of Europe and the 
Stereomithes to which that family has been referred, and 
concludes that though it is not impossible that some affinity 
between them may exist, its nature is quite uncertain. 

In a notice of the same memoir (15) the present writer 
has compared the skeleton of Phororhacos with several 
other types, and a considerable degree of resemblance with 
the Cariama (Dickolop/ms) was found to exist, particularly 
in the structure of the metatarsus. If further investigation 
of the specimens themselves should confirm these observa- 
tions, the Cariama would appear to be related to these 
gigantic and highly specialised extinct birds somewhat as 
the recent Armadillos are to the extinct Glyptodonts. In 
both cases the recent forms cannot be regarded as direct 
descendants of the fossil giants, but rather as more gene- 
ralised descendants from the same common stock, which 
have escaped extinction both on account of their smaller 
size, and more particularly, because being less specialised 
they were less affected by changes in the conditions of life. 

The specimens described by Ameghino have been pur- 
chased by the Trustees of the British Museum, and many 
of them may now be seen at the Natural History Museum. 

No papers of importance dealing with upper tertiary 
birds have appeared within the time to which this review- 
is limited. 

Quaternary Birds. — During the last three years some 
important additions to our knowledge of the extinct stru- 
thious birds of Madagascar, the sEpyornitJiida!, have been 
made. Until 1893 only the bones of the hind limb and 
some imperfect vertebrae of these birds were known, and 
no paper describing new material had appeared since the 
publication of Milne Edwards and Grandidier's classical 
memoir in 1870. In 1893 Burckhardt (16) gave a very 
detailed account of a small collection of /Epyornis remains 
that had been obtained at Sirabe. in Central Madagascar. 
This included not only limb bones and vertebrae, but also 
the greater part of the pelvis and sacrum ; all the specimens 
were referred to a new species, JE. Hildebrandti, which the 
author compares with those previously known and with the 


other RatitcE. The conclusions he arrives at are of con- 
siderable interest. Milne Edwards and Grandidier expressed 
an opinion that JEpyornis is related to the Dinornit hides, 
coming between that family and the Australasian Ratites, 
Casuarius and Dromceus, the latter of which according to 
some writers is the most primitive of the group. Burck- 
hardt also considers that JEpyornis is most closely related 
to Casuarius and Dromczus, but believes that the resem- 
blances between it and Dinomis are merely the result of 
parallelism in evolution, the skeleton in both cases having 
become extremely massive. On the other hand, he believes 
that in some of the characters of the pelvis and other parts 
of the skeleton, and also in the structure of the egg-shell, 
jEpyornis approaches Struthio, and suggests that from the 
primitive Dromczus-Casuarius stock the Dinornithidce and 
Apteryx were descended on the east, while towards the 
west a branch arose which split up into the /Epyornithidce 
and Strut hionidcz. This view, though it may perhaps ap- 
pear to be supported by the geographical distribution of the 
families concerned, cannot be regarded as established. The 
structure of the skull and shoulder-oirdle and sternum when 
known will probably settle this question. 

At the beginning of 1894 the present writer described 
{17, 18) a species of yEpyornis, /E. Titan, far larger than 
any then known. The tibia is about thirty-one inches long, 
enormously massive, even more so than that of Pachyornis 
elephantopus. In January of the same year Milne Edwards 
and Grandidier (19) published a preliminary notice of a very 
large collection of AEpyornis remains. They name some four 
or five new species of sEpyornis, and establish a new genus, 
Mullerornis, for the reception of three smaller forms of 
more slender build than Aipyornis. A large part of the 
skeleton of one of the new species is very briefly described. 
The skull is said to be less flattened than that of Dinomis, 
and at the same time narrower and longer ; the brain was 
proportionately considerably larger. The mandible some- 
what resembles that of Rhea, while the sternum ap- 
proaches that of Apteryx in structure. The coraco- 
scapula is small, and bears a shallow glenoid cavity for the 


head of the rudimentary humerus. Further descriptions 
and figures of this valuable specimen will no doubt be of 
great service in settling the question of the affinities of the 
family. The authors incline strongly to the view that 
sEpyornis is closely related to Dinornis, and, as in their 
former paper on this subject, suggest the former existence 
of a land connection between Madagascar and New Zealand 
to account for this relationship. In conclusion they state 
that there is clear evidence that Aipyornis was contemporary 
with man, and also mention that remains of a species of 
Aphanapteryx and a large extinct anserine bird occur in the 
same deposits. 

In a later paper (20) the same authors describe in some 
detail the skull of one of the smaller forms included in the 
genus Mullerornis. This is said to differ widely from that 
of JEpyornis, the cranial region being much less depressed 
and the frontals raised so as to form a prominent boss. The 
basi-pterygoid processes are only slightly developed, and the 
anterior region of the premaxillae is more compressed and 
forms a rounded keel above. Of all recent Struthious birds 
the Cassowary is said to most resemble Mullerornis, both in 
its cranial characters and in many points in the remainder 
of the skeleton. 

An interesting account of the mode of occurrence of the 
bones and eggs of the JEpyornithida is given by Mr. J. T. 
Last (21), who resided in the island for some time and made 
collections in several localities. It appears that the bones 
are usually found in the dried beds of ancient lakes or in 
swamps, where they sometimes occur in large numbers ; 
the eggs, on the other hand, are rarely found in such 
places, but occur in great quantities (in fragments) in the 
shifting sand-dunes round the coast. 

In 1893 Professor Jeffrey Parker (22) published a new 
classification of the Moas, founded on the characters of the 
skulls. This paper is merely an abstract from his important 
memoir on the cranial osteology of the group, which will be 
noticed below. Here we need only mention that the 
Dinornit hidce were subdivided into three sub-families, the 
Dinormthince, Anomalopterygince, and the Emeince, and 


that of all the genera Mesopteryx is considered to be the 
least specialised, and retains most nearly the ancestral 
characters of the family. At the same time it was also 
shown that some species probably possessed a frontal 
crest of large feathers, the points of insertion of which 
are marked by a series of pits on the cranial surface ; in 
some cases this character seems to have been a sexual one. 

In the same year Hutton (23) published a paper which 
may be regarded as an appendix to his important memoir, 
"On the Moas of New Zealand," which appeared in 1892, 
and consequently does not fall within the scope of the 
present review. In this appendix the author states that in 
his opinion it is necessary to subdivide the various genera 
of the Dinornithidae into more species than had hitherto 
been done ; since it is only by keeping the various species 
and varieties distinct, that the relative ages of the various 
superficial deposits in which their remains occur can be 
ascertained. The method of subdivision employed by 
him seems, however, to be open to the objection that it 
is an extremely arbitrary and artificial one, for in his former 
paper above referred to, as well as in the present one, he 
relies mainly, and in many cases entirely, on measurements 
of the long bones for separating the species. When we 
consider that it is possible to trace an almost complete 
gradation in size between the larger and smaller specimens 
of any given bone, it is clear that the number of species 
into which the series is divided, will depend upon personal 
opinion as to the latitude to be allowed for individual 
variation. In some cases where small differences in size, 
accompanied by other slight variations, are constant in two 
forms from different localities, the careful records and 
measurements given by Professor Hutton are of much 
interest and importance, but even in such cases it seems 
better to regard such small differences as indicating local 
races rather than distinct species. 

Dr. H. O. Forbes (24) has severely criticised Professor 
Hutton's methods, and points out that in some cases the 
measurements given for one species fall within the limits 
assigned to another. 


Casts of a number of pieces of limb bones of a small 
Moa, Anomalopteryx antiqua, which were discovered be- 
neath a lava-flow at Timaru, are described and figured by 
Hutton (25). These specimens were first noticed by 
Forbes, who states that they were accompanied by re- 
mains of Apteryx. As to the age of the deposits in which 
these fragments occur there is much difference of opinion, 
and they have been successively referred to the Eocene, 
Miocene, Pliocene, and Pleistocene. Forbes believes 
that they are Pleistocene or at latest Upper Pliocene, 
while Hutton regards them as Miocene or Pliocene. In 
any case the specimens, which have been lost, were so 
imperfect that conclusions dependent upon them must be 
received with caution. 

In a subsequent memoir (26) by the same author the 
structure of the axial skeleton in the various a-enera is dis- 
cussed, and the descriptions of the various forms of pelvis 
and sterna are very useful, as also are the references to the 
published figures of various portions of the axial skeleton 
of the different forms. 

Some ten years ago De Vis announced the discovery 
in Queensland of a femur of a species of Dinornis. The 
occurrence of the New Zealand type of Ratite bird in 
Australia would, of course, be a matter of great interest 
and importance in questions relating to the geological 
history of two areas ; but the great difficulty in accurately 
determining isolated bird bones made it seem probable that 
in this case a mistake had been made. This suspicion 
would appear to be well founded, for Hutton, having 
had an opportunity of examining casts of the type speci- 
men of the so-called Dinornis Queenslandice, states (27) 
that it differs widely from all Dinornithine femora with 
which he is acquainted. He considers that the bone is 
that of a bird related to Dromseus (the Emu), and coming 
between that genus and Dromornis, an extinct Australian 
form described by Owen. 

The most valuable contribution to our knowledge of the 
morphology of the Moas that has been published for many 
years is Professor Jeffrey Parker's paper (28) on the cranial 


osteology of the group. He has had opportunities of exa- 
mining a very large number of skulls, some of which are 
those of young individuals in which the sutures are still 
open, and has, therefore, been able to give a very detailed 
account of the structure of this, the most important portion 
of the skeleton. Moreover, he has given a scheme of 
classification of the group, founded exclusively on cranial 
characters ; the importance of this is obvious to any one 
acquainted with the terrible state of confusion into which, for 
various reasons, the nomenclature of the Moas has got. Five 
genera are recognised, and it would be very advantageous 
if these could finally be adopted, particularly as they agree 
in the main with those accepted by Lydekker in his cata- 
logue of the British Museum collection, which, containing 
as it does the types of most of the species, must be the 
final court of appeal in most questions relating to the 
nomenclature of the family. Professor Parker has added 
a very detailed comparison of the Dinornithine skull with 
those of the other Struthious birds, and arrives at some inter- 
esting results as to the relationships existing between the 
various types. He considers that the Ratitae are a poly- 
phyletic group, Rhea and Struthio having originated inde- 
pendently of one another and of the forms inhabiting 
Australia and New Zealand. The latter arose from a 
common stock which early divided into two branches, 
the one giving rise to the Australian genera, Dromczus 
and Casnarius, the other to the New Zealand forms. The 
latter again divided into two branches, one leading to the 
ApterygidcB, the other to the Dinornithidce. Of this family 
Dinoruis and Emeus are regarded as having diverged most 
widely from the ancestral type, which is probably most 
nearly represented by Mesopteryx. This view differs from 
that of Burckhardt mainly in the refusal to admit a common 
ancestry of Struthio and the Casuariidae, otherwise it is in 
general agreement with it, and is supported by the geo- 
graphical distribution of the various forms. Unfortunately 
palaeontology throws little or no light on the history of the 
Struthious birds, no fossil form that can be referred to that 
group with certainty being known from strata older than 


the Pliocene. It is true that many extinct birds, as, for 
example, the Gast omit hides and Stereornitkes, have been 
referred to it, but in no case does it appear probable that 
we have to do with either actual ancestors or even offshoots 
of the ancestral Struthious stock. So far, therefore, as palae- 
ontology is concerned we have no means as yet of deter- 
mining the relations of the Ratitse either with one another 
or with other birds, and it is on such studies of the com- 
parative anatomy of the various groups, as that given by 
Professor Parker in the case of the skull, that we must rely 
for information on this point. 

Numerous papers have appeared lately dealing with the 
vexed question of the date of extinction of the Moas, and 
the points of view from which the problem has been attacked 
are very numerous. On the whole the evidence brought 
forward seems in favour of the view, so ably advocated by 
Dr. H. O. Forbes and others, that these birds have died out 
comparatively recently, and that their extinction is mainly due 
to the persecution they suffered from the Maoris, who hunted 
them down for food, and probably also destroyed their 
eggs. One of the reasons for believing that they survived 
till quite lately is the occurrence of portions of their bodies 
with dried flesh and feathers still adhering, several additional 
instances of which have been brought to light during the 
last year or so. Hamilton (29) has given a very interesting 
account of the various finds of Moa feathers, and more par- 
ticularly of one which he himself investigated. In this case 
a large quantity of feathers, probably belonging to a species 
of Megalapteryx, were found in a cavern near the head of 
the River Waikaia, where a leg of the same bird with the 
flesh and skin still adherent had previously been discovered. 

Some important discoveries of remains of extinct birds 
other than the Moas have been recently made in the New 
Zealand region. In a fissure in the limestone at Castle 
Rocks, Southland, Hamilton found an immense quantity of 
the bones of birds which appear to have fallen into the 
opening as into a pit-fall ; though this can hardly have been 
the case with the large extinct eagles, Harpagornis, remains 
of both species of which occur. The remainder are nearly 


all flightless forms, including Anomalopteryx, a large species 
Fulica, much like that found in the Chatham Islands (see 
below), a small Weka-rail, Aptoruis, Notornis, and several 
others. An account of these, together with elaborate tables 
of measurements of the limb bones of some of them, will be 
found in Hamilton's paper (30). 

The most important of all the recent discoveries in this 
region is, without doubt, that made by Dr. H. O. Forbes. 
In 1892 (32) he announced in Nature that he had received 
from the Chatham Islands (about 500 miles east of New 
Zealand) a skull of a large rail closely resembling the 
extinct Aphanapteryx of Mauritius ; to this the name Apha- 
napteryx Hazvkinsi was given. A large collection of bird 
remains, subsequently obtained from the same locality, 
contained all the more important bones of many individuals 
not only of this species, but also of several other extinct 
forms. Among the more notable of these were a large 
Coot, Fulica chathamica, very similar to the Mauritian 
species, F. Newtoni ; a new type of Crow, Palccocorax 
moriorum, said to be most nearly related to the Gymnorhine 
group ; an extinct Swan, Chenopis, besides several other 
species, most of which are still inhabitants of the Islands. 
Several of the extinct forms have not yet been described, 
but of Aphanapteryx Hawkinsi and Palcsocorax moriorum 
a short account was published in the Ibis. At the same 
time a new genus, DiapJiorapteryx, was established for the 
reception of the former species. Subsequently, however, 
the new name was withdrawn, and Forbes expressed his 
conviction that the Chatham Island and Mauritius birds are 
not generically distinct, and must, therefore, both be referred 
to Aphanapteryx. This opinion he defends in a short paper 
(34), illustrated by figures of the humerus, sternum, and 
premaxillse of the two forms. 

In a paper by the present writer (35), on the osteology 
of the Chatham Island bird, a number of differences between 
it and Aphanapteryx broecki are pointed out ; and some of 
them, as, for example, the great dissimilarity between the 
metatarsi, are clearly of generic value, so that the name 
Diaphorapteryx was again adopted. 


The assumed generic identity of these two forms was the 
most important new evidence brought forward by Forbes in 
his paper supporting the hypothesis of the former existence 
of an Antarctic Continent ; but in the paper just referred to 
(35) it was shown that, as far as the birds are concerned, 
there is no evidence that the Chatham Islands have been 
united with any land area, and that the presence of two 
similar flightless rails on two islands remote from one 
another is no proof of any former land connection between 
them. In such a case it seems far more reasonable to sup- 
pose that both the islands may have been colonised by the 
same or allied forms of flying rails which have subsequently 
lost their powers of flight, owing to the very fact of their 
insular conditions of life. An instance of this on a smaller 
scale is found in the case of Tristan d'Acunha and Gouo-h 
Islands, which are about 200 miles from one another and 
about the same distance from the Cape of Good Hope. 
Each of these islands is inhabited by a distinct species 
of Gallinule (Porphyriornis), which closely resemble one 
another and are incapable of flight ; yet no one has sug- 
gested that on that account these islands were formerly 
united by land, either with one another or to Africa. 

It is a fortunate coincidence that while the relationship 
between Diaphorapteryx and Aphanapteryx was still in 
dispute some additional remains of the latter were described. 
These bones, together with those of many other species, 
including the Dodo, Lophopsittacus mauritianus, Fulica 
Newtoni, etc., were described by Newton and Gadow in 
a well-illustrated memoir (36). Besides adding much to 
our knowledge of previously known extinct birds, the 
authors have been able to describe a number of new 
ones. They have also published a figure of the restored 
skeleton of the Dodo, which in several respects is more 
correct than those which have previously appeared. The 
whole of the remains described were obtained from the 
Mare aux Songes, from which previously a large quantity 
of Dodo bones had been collected. Besides the bones of 
birds those of the large extinct lizard, Didosaurus, and 
carapaces of Tortoises were found. 



(1) MARSH, O. C. A new Cretaceous Bird allied to Hesperornis. 

Anier. Journ. Science, vol. xlv., p. 81, 1893. 

(2) Marsh, O. C. On a Gigantic Bird from the Eocene of New 

Jersey. Amer. Journ. Science, vol. xlviii., p. 344, 1894. 

(3) Cope, E. D. On Cyphornis, an Extinct Genus of Birds. 

Journ. Acad. Nat. Sci., Philadelphia, vol. ix., p. 449. 

(4) Edwards, A. Milne. Sur les Oiseaux fossiles des Depots 

Eocene de Phosphate de Chaux du Sud de la France. 
Comptes Rendu s. Congrcs Omithologique International, Buda 
Pesth, pt. ii. (Partie Scientijique), p. 60, 1892. 

(5) Lydekker, R. L. On some Bird Bones from the Miocene of 

Grive-St.-Alban, Department of Isere, France. Proc. Zool. 
Soc, p. 517, 1893. 

(6) AMEGHINO, F. Enumeracion systematica de las especias de 

mamijeros Josiles colec. por C. Anicghino en los terrenos eocenos 
de la Patagonia austral., p. 24, 1887. 

(7) Ameghino, F. Aves fosiles argentinas. Revista Argentina 

de Historia Natural, tome i., p. 255, 1891. 

(8) Moreno and Mercerat. Catalogo de los Pajaros fosiles de 

la Republica Argentina. Palceontologia Argentina, tome i. 
(Anales del Museo de la Plata), 1891. 

(9) Ameghino, F. Enumeracion de los aves fosiles de la Repub- 

lica Argentina. Revista Arg. Hist. Nat., tome i., p. 441, 1891. 

(10) GADOW, H. Bronn's Thierreich. Aves, tome ii. (Systemati- 

scher Theil), p. 106. 

(11) Lydekker, R. L. On the Extinct Giant Birds of Argentina. 

Ibis, p. 40, 1893. 

(12) Ameghino, F. Sur les Oiseaux fossiles de Patagonie. Boletin 

del Instituto Geografico Argentino, tomexv., cap. 1 1 and 12, 1895. 

(13) Lydekker, R. L. The La Plata Museum. Natural Science, 

vol. iv., p. 126, 1894. 

(14) Lydekker, R. L. The Giant Birds of South America. Know- 

ledge, vol. xviii., p. 125, 1895. 

(15) Andrews, C.W. Remarks on the Stereornithes. Ibis, p. 1 , 1 896. 

(16) Burckhardt, R. Ueber ^pyornis. Palczontologische Ab- 

haudluugen, Bd. ii. (Neue Folgc), p. 127, 1893. 

(17) Andrews, C. W. Note on a New Species of ^Epyornis {/E. 

Titan). Geological Magazine, 4th Dec, vol. i., p. 18, 1894. 

(18) Andrews, C. W. On some Remains of yEpyornis in the 

British Museum. Proc. Zool. Soc, p. 108, 1894. 

(19) Milne Edwards and Grandidier. Observations sur les 

^pyornis de Madagascar. Comptes Rendus Acad. Sci., t. 
cxviii., p. 122, 1894. 



(20) Milne Edwards and Grandidier. Sur des Ossements 

d'Oiseaux provenant des Terrains Recents de Madagascar. 
Bulletin du Museum d ' Histoire Naturelle, p. 9, 1895. 

(21) LAST, J. T. On the Bones of ^Epyornis, and on the Localities 

and Conditions in which they are found. Proc. Zool. Soc., 
p. 123, 1894. 

(22) Parker, T. J. On the Classification and Mutual Relations 

of the Dinomithidce. Trans. New Zealand Instil., vol. xxv., 
p. 1, 1893. 

(23) HUTTON, F. W. New Species of Moas. Tow. eit., p. 6. 

(24) Forbes, H. O. The Moas of New Zealand. Natural Science, 

vol. ii., p. 374, 1893. 

(25) HUTTON, F. W. On Anomalopteryx antiqua. Trans. N. Z. 

Instit., vol. xxv., p. 14. 

(26) HUTTON, F. W. On the Axial Skeleton in the DinornithidcB. 

Trans. N. Z. Instit., vol. xxvii., p. 157, 1895. 

(27) HUTTON, F. W. On Dinomis (?) QueenslandicB. Proc. Linn. 

Soc, New South Wales, vol. viii. (2nd series), p. 7, 1893. 

(28) Parker, T. J. On the Cranial Osteology, Classification, and 

Phylogeny of the DinornithidcB. Trans. Zool. Soc, vol. xiii., 

P- 373, i895. 

(29) Hamilton, A. On the Feathers of a Small Moa. Trans. N. 

Z. Instit., vol. xxvii., 1895, p. 232, 1894. 

(30) HAMILTON, A. On the Fissures and Caves at the Castle 

Rocks, Southland ; with a Description of the Remains of 
the Existing and Extinct Birds found in them. Trans. N. Z. 
Instit., vol. xxv., p. 88. 

(31) HAMILTON, A. Result of a Further Exploration of the Bone 

Fissure at Castle Rocks, Southland. Trans. N. Z. Instit., 
vol. xxvi., p. 226. 

(32) Forbes, H. O. On a Recent Discovery of the Remains of 

Extinct Birds in New Zealand. Nature, vol. xlv., p. 416, 

(33) Forbes, H. O. Ibis, p. 253, 1893. 

(34) FORBES, H. O. On the Aphanapteryx of Mauritius and of the 

Chatham Islands. Ann. Mag. Nat. Hist., ser. 6, vol. xii., p. 
65, 1893. 

(35) Andrews, C. W. On the Extinct Birds of the Chatham Is- 

lands. I. The Osteology of Diaphorapteryx Hawkinsi. 
Zoologies Novitates, vol. iii., p. J^, Tring, 1896. 

(36) Newton, E., and Gadovv, H. On additional Bones of the 

Dodo and other Extinct Birds of Mauritius, obtained by M. 
Theodore Sauzier. Trans. Zool. Soc, vol. xiii., p. 281, 1893. 

C. W. Andrews. 

Science |3ragre 


No. 30. August, 1896. Vol. V. 



Continued from vol. Hi., p. 1S5. 

WHEN we come to consider how to imagine the 
mode by which light discharges an electrified 
surface, one of the first hypotheses is that it may be by a 
kind of proof-plane action, the illuminated surface being 
disintegrated and its charged molecules evaporated away, 
taking their charges with them. 

The first objection to such a hypothesis is that the dis- 
integrating action of light ought to be otherwise perceptible 
either to microscopic inspection or to a delicate balance 
which should determine the loss of material. 

It has, however, often been suspected that metals may 
evaporate more or less, and the fact of their smell seems to 
establish the fact, so it may be well to consider how small 
a loss of material will serve to explain the observed loss of 

If we assume that each molecule so evaporated has the 
ionic charge on one of its atoms reversed, or, more simply, 
if we assume that each atom carries off a quantity of 
electricity of the order io -11 electrostatic unit, its maximum 
possible and customary value, then the amount of electricity 
associated with the one gramme of evaporated silver is 
900 contants or 3 x io 12 electrodal units. 



Now a silver plate 14 centimetres square, under certain 
conditions of arc illumination, was found in the writer's 
laboratory to lose negative electricity at the rate of 30 
electrostatic units per minute when kept electrified to 80 
volts in fairly free space. Hence the time that would 
elapse before the above plate would lose a tenth of a 
milligramme of its substance is ten million minutes or nearly 
a century. 

Such a ratio of loss as that could not be detected by a 
balance, even in the case of silver, which is the substance 
most suitable for detecting a small electrolytic loss by 
weight. But now suppose that the discharge is not of so 
atomic a character, but that little flakes or pieces of the 
metal are driven off under the electric stress, so that the 
charge per gramme lost is very much less. 

In that case the disintegration of surface might be 
perceived, but there are many difficulties in the way of 
supposing such an action. 

The electric tension even when on the verge of dis- 
ruption, when the surface is charged to many thousand 
volts, is by no means comparable to the forces of cohesion. 

And the action of light occurs at so low a tension that 
it is impossible that its action is a mere bringing down of 
the limit at which disruptive discharge begins. 

The action of light is much more like a quiet atomic or 
molecular process than it is like a disruptive discharge from 
the substance in bulk. 

It may, however, be worth noticing that the electric re- 
pulsive force experienced by an atom when 011 a surface 
charged to the disruptive limit is not incomparably less than 
the average force of cohesion acting on such an atom. The 
tenacity of a metal may be taken as io 9 cgs. units, or about 
icr 7 degree, per superficial molecule. The electric force 
acting on an atom in a potential gradient of 30,000 volts per 
centimetre, which is the disruptive limit under ordinary 
atmospheric conditions, is about io -9 degrees — one-hun- 
dredth of the average cohesive force ; so it would not be 
unduly speculative to conceive it possible that circumstances 
connected with heat and other motors should occasionally 


render individual atoms detachable under stresses approach- 
ing so near to the average limit, and this would be one way 
of representing disruptive discharge. 

Against this, however, must be set the fact that the dis- 
ruptive limit depends greatly on the atmospheric condi- 
tions, on the pressure and nature of the gas in contact with 
the metal ; therefore it would appear that even for disrup- 
tive discharge we must look to an interaction between the 
molecule of the metal and that of the medium in contact 
with it, rather than to a simple disruption of the metal 

What is certain is that the charge is carried away by 
particles (atoms or otherwise) which travel along the lines 
of force to the oppositely electrified surfaces. It may con- 
ceivably be that the conveyers of charge are the electrons 
themselves ; in other words, that the negative ends of the 
lines of force are detached from the charged body under 
the action of light, and that the line therefore promptly 
shuts up. It is more probable, however, there are no such 
detached electrons or atomic charges divorced from matter, 
but that the negative charge is conveyed by material atoms, 
whether they be the atoms of the metal or of the sur- 
rounding gas. To examine the question whether the con- 
veying atoms belonged to the metal or to the gas, a number 
of experiments have been made in my laboratory with the 
object of testing the presence of metallic particles or vapour 
near an electrified metal rapidly discharging under the 
action of light. 

The metals most easy to detect in small quantities are 
in general perhaps silver, iron and sodium. Silver, by its 
reflecting power when deposited upon glass ; iron, by its 
magnetic properties ; and sodium, by the light it causes a 
non-luminous flame to emit. Silver plates, with their clean 
edges opposed to the surface of plate glass, were oppositely 
electrified so that any charge given off from the silver edge 
should be deposited upon the glass as upon the dielectric of 
a Leyden jar, and were kept thus strongly illuminated by 
an arc light for hours ; the glass was then examined for 
transparency. A decided deposit was found near the illu- 


minated region, but there appeared nothing metallic about 
it, and it was easily dusted off. It seemed to be merely 
dust out of the air. So the experiment was repeated in a 
dust-free chamber, containing air filtered slowly through 
Ions' tubes of cotton-wool, and now not the faintest local 
dimming of the surface could be observed, although the illu- 
mination and electrification lasted for days. So the answer 
for silver was in the negative. 

Next a non-magnetic substance was hung in a powerful 
converging magnetic field in the neighbourhood of clean 
illuminated and oppositely electrified iron, to see if by con- 
densation of evaporated iron, it was possible that it became 
magnetic. A minute torsion-bar of copper suspended over 
a clean, conical, vertically pointing electro-magnet's pole 
was the best arrangement. There were difficulties about 
this experiment on account of electrodal and other forces, 
but so far as disturbance could be eliminated the result 
for iron was also negative. 

Then the most elaborate series of observation was made 
on metallic sodium kept in an atmosphere of highly purified 
hydrogen, the gas being supplied through a long series of 
drying tubes, and kept burning as a small jet just after it 
had passed over the sodium surface. By a mechanical 
arrangement the sodium could be cut to a clean surface from 
outside, and when the gas was pure this surface lasted a 
fairly long time, and under illumination it discharged elec- 
tricity supplied by several dry piles in series, so that a 
considerable supply of electricity could be drawn from the 
flame whenever light from an arc lamp was allowed to 
fall on the sodium surface through a quartz window. The 
flame was looked at either direct or through a small 
spectroscope, and though the sodium line could not be kept 
wholly absent, its occasional presence depended in no way 
on whether the surface was positively or negatively or not 
at all electrified, nor on whether the light was or was not 
shining on it. 

Hence I conclude that the discharge of electricity from 
illuminated surfaces is not effected by evaporation of those 
surfaces, but that the molecules which convey the charge 


belono- to something in the gas, and not to the illuminated 

It may be asked whether dust in the air has any part in 
the action, but, so far as I can find, it has none at low 
tensions. The discharge rate from the silver surfaces, for 
instance, was just about as rapid in a dust-free atmosphere 
as when dust was present. 

The proof that the discharge is effected by molecules of 
some kind, or at least by something which travels along the 
lines of electrostatic force was given by Righi. He elec- 
trified a small metallic cylinder of which only one generating 
line was free from varnish, and therefore clean enough to dis- 
charge electricity. This cylinder being negatively electrified 
in front of an earthed plate, an exploring terminal of an 
electroscope could ascertain which part of the plate was 
receiving a charge, as the cylinder was rotated on its axis, 
a movable slit being arranged in the plate for this purpose ; 
and it was found to be always near one extremity of a circular 
arc of which the discharging line constituted the other 
extremity. He further found that if the illuminated body 
were free to move it receded like an electric windmill, 
proving that it had imparted its charge to something 
possessing appreciable inertia. 

The inertia of the gaseous particles would indeed cause 
some divergence from the above circular orbits in which 
the electrical force is urging them, but the force is so great 
and the mass is so small that the deviation is not noticeable. 
Moreover, the charged atom has to make its way among a 
crowd of others by a process very similar to what occurs in 
electrolysis, so that the path of the electric charge follows 
almost accurately the line of electric force. In that sense it 
may be said to represent the motion of an electron or free 
electric charge, without committing the speaker to the 
hypothesis that such charges divorced from their usual 
boundary conditions on matter can really exist. 

If a gaseous atom can receive a charge from an elec- 
trified surface there is no difficulty in understanding what 
it does with it, nor how, by such a process, the electrified 
body gets discharged, but the difficulty is to realise how an 


atom can so receive a charge. Under ordinary circum- 
stances it is certain that gas molecules cannot acquire a 
charge until the electrical tension rises to the disruptive 
point; but there is a certain condition into which a gas can 
be thrown, similar, if not identical, with that which chemists 
speak of as dissociation, wherein a gas becomes a conductor, 
that is to say, its particles do really act as carriers of electric 
charges, and may be spoken of as detached and specifically 
charged atoms. 

Now in a vacuum tube, w T e learn experimentally from 
Mr. Crooks, that at high vacua the negatively charged atoms 
are vigorously repelled from a negative electrode, and, 
shooting out from it in straight lines, constitute what are 
known as cathode rays. It appears as if the electric 
discharge itself were carried on in a vacuum tube by a quiet, 
imperceptible, electrolytic action, originating at the anode or 
positive electrode, that this discharge fills the whole tube 
with positive electrification up to within a short distance of 
the cathode. In this short distance there is accordingly a 
steep potential gradient, and any stray negative atoms 
finding themselves therein are shot out of it with immense 
velocity, and constitute what are called cathode rays. 
Some doubt has been felt as to the essential nature of 
cathode rays, but there is hardly any good reason for the 
belief that they are anything else than a stream of negatively 
charged atoms of matter. They need not have recently 
received a charge, their charge may be intrinsic ; what we 
observe is their repulsion, not as if guided through a resist- 
ing medium by electric force, but as if propelled violently 
inside a thinned layer and left to the first law of motion 

Great interest has been felt in this cathode stream for 
a quarter of a century, but within the present year its im- 
portance has become immense owing to the discovery of 
Rontgen that a surface on which the stream impinges be- 
comes capable of emitting a novel kind of radiation which 
travels even more persistently in straight lines, and is not 
readily stopped by material obstacles. This discovery 
must ultimately throw a great deal of light upon the whole 


subject, and it is over soon to attempt to forecast its 
probable development ; nevertheless a partial attempt may 
be made for what it is worth. 

The new radiation appears to differ from ordinary 
ultra-violet radiation only in the matter of wave-length. 
Its wave-length is probably extremely short, not vastly 
greater than the size of atoms, and all its other known 
properties and peculiarities will follow from that according 
to known theories of dispersion, especially the electro- 
magnetic one of von Helmholtz. 

Now this X radiation, when it falls upon an electrified 
surface, discharges it, somewhat in the same fashion that 
ultra-violet light does ; but whereas light discharges 
electricity solely, or at any rate chiefly, of the negative sign, 
this X radiation discharges both positive and negative ; 
and indeed it seems to act by converting the gas or other 
insulating material near a charged body into a conductor. 
This it probably does by dissociating the substance into 
charged atoms which are then free to act as carriers, and 
speedily convey to a distance the charge of the electrified 
body by journeys along the lines of force. 

It may be that ultra-violet light acts in somewhat the 
same way, but not in exactly the same way. The air is 
transparent to ultra-violet light, it is not perfectly trans- 
parent to X rays. 

There is no difficulty in supposing that the X rays 
dissociate some ingredient of the atmosphere, but there is 
great difficulty in supposing ordinary ultra-violet light to be 
able to do so. What the ultra-violet light chiefly does is 
to promote or to create the conditions necessary for the 
ready interchange of electric charge between gas and 
solid ; and that this is so is practically proved by the great 
importance of the nature of the solid surface, as well as of 
the gas in contact with it. The gas seems indeed of 
secondary importance, but the cleanness and oxiclisability 
of the solid is essential to a rapid and ready discharge with 
ordinary light from the visible spectrum. High ultra- 
violet light can act indeed over a wider range, and where- 
as light of long wave-length can only discharge negative 


electricity, it is probable that light of extremely short wave- 
length can discharge positive also, and from surfaces not 
specially clean nor oxidisable. 

The X rays seem to go farther in the same direction ; 
that is to say, their activity does not appear to depend much 
upon the nature of the surface, nor do they seem to discrimi- 
nate much between positive and negative electrification. 

We may surmise, then, that long-wave light is effective 
in promoting discharge only when dissociated or incipiently 
dissociated atoms are already present in the neighbourhood 
of the surface. It is otherwise known that strongly electro- 
positive substances, like clean sodium or zinc, are surrounded 
by a number of electro-negative (chiefly oxygen) atoms, 
straining to get at it. And, similarly, a negatively charged 
surface may be surrounded by a number of straining posi- 
tive atoms. Under these circumstances it is not difficult to 
picture the result of impinging waves of light, and of the 
electrical oscillations which they must necessarily set up, as 
resulting in an interchange of electricity between the surface 
and the gas which otherwise might not have occurred. 

When positive electricity has thus been received by the 
metal from the air under the action of light, detached nega- 
tive ions will be left in the atmosphere, and these will be 
repelled by the body if kept negatively electrified, and so 
may constitute a kind of feeble cathode ray. 

Thus it appears as if there were a sort of reciprocal 
action ; the impact of light on a negatively electrified sur- 
face results in the production of something akin to cathode 
rays, and the impact of cathode rays upon a positively elec- 
trified surface results in something akin to light. 

Another instance of reciprocity has also been observed. 
Certain substances exposed to X rays fluoresce strongly, 
that is, emit light which in some cases persists an appreciable 
time, and some of these substances when made to fluoresce 
by exposure to light begin to emit X rays and continue to 
emit them for long after, as has been observed by M. 

There is one matter dealt with in the last article which 
requires more cautious handling than it then received, and 


that is the discharge of positive electricity — i.e., the recep- 
tion of negative by certain substances. It is a phenomenon 
which undoubtedly occurs as an experimental fact, but if 
we proceed to look into the cause of it we find its detailed 
character by no means so obvious Certainly it depends a 
great deal on the surroundings, and there is reason to be- 
lieve that if a positively charged body were surrounded by 
a surface incapable of giving off negative electricity, then 
the apparent discharge of positive might not occur. 

The question is complicated by the simple facts (a) that 
we cannot have a charged body without an equal opposite 
charge on surfaces opposed to it, and (6) that every sur- 
face reflects and scatters some of the incident light which 
therefore partly falls upon the oppositely electrified sur- 
face. Hence when a positively charged body loses its 
charge, it may be not through a direct action of the light 
upon itself, but by reason of the action of the reflected 
and scattered light on the negatively electrified surfaces in 
its neighbourhood. 

On this hypothesis a surface which appears to lose 
positive more quickly that negative is one which of itself 
hardly loses any electricity at all ; it loses negative slowly 
but it is exposed to surfaces which can emit it more quickly, 
and hence when it is positively electrified and they are 
inductively negative, it receives from them a negative 
charge more rapidly than it was able to give one out. 

A large number of exoeriments have been made in the 

O J. 

writer's laboratory to test this point, mostly by means of 
regular reflectors so as to avoid scattered light as far as 
possible ; the details are somewhat technical and trouble- 
some, and the verydust of the air is apt to scatter a good 
deal of active light ; but the result is, on the whole, to 
substantiate the above-mentioned idea, which also possesses 
the powerful support of Messrs. Elster and Geitel, that the 
loss of positive electricity under the action of light is an 
indirect and secondary phenomenon. 

It appears, however, that under X rays both points of 
electricity are discharged equally, and if these X rays are, 
as everything now indicates them to be, an extension into 



very much higher regions of the spectrum of transverse 
ethereal vibrations, then it must become a question of degree 
and of wave length, as implied above by the writer, and no 
perfectly simple statement can be made. 

The activity of ordinary sunlight in promoting the 
discharge of electricity into the atmosphere is evidently a 
question of great meteorological importance ; but it is enor- 
mously affected by the condition of the earth's atmosphere, 
At high elevations the rays are very active, but in valleys 
the power is less, and on many days in a town there is 
hardly any power left at all. The writer's assistant, Mr. 
Davies, constructed a portable apparatus with which he 
made many observations in Wales and other places during 
last summer at different heights and at different periods of the 
day. The results are such as might naturally have been 
expected, but we do not yet know whether the sun 
emits any X rays at all detectable in the higher region of the 
atmosphere, or whether this latter variety of radiation is 
an artificial product recently introduced by man into the 
operations of Nature. 

Oliver Lodge. 



IN two previous articles l some account has been given of 
the genus Sphenophyllum, with special reference to the 
structure of the strobilus. I now propose to add a brief 
summary of our knowledge of this interesting type of 
extinct plants, which has been fully dealt with by William- 
son and Scott in their memoir on Catamites, Calamostachys, 
and Sphenophyllum} 

Every collector of Coal-Measure plants must be 
familiar with the fragments of slender stems bearing 
regular whorls of wedge-shaped leaves, which are fre- 
quently found in the Upper Carboniferous shales, or in the 
ironstone nodules of Coalbrookdale and other places. 
Writing in 1822, Brongniart 3 describes and figures a 
well-preserved impression of a species of Sphenophyllum 
under the name Sphenophyllitcs cmarginatus, and speaks 
of it as a plant without any living generic analogue. In 
the classic Prodrome dune histoire des vegUazix fossiles, the 
same author eives the following definition of this fossil 
genus, and adopts the generic name Sphenophyllum i : — 

" Tige simple, articulee ; feuilles verticillees, au nombre 
de six a douze, distinctes jusqua leur base, cuneiformes, 
entieres ou emarginees, ou meme bifides, a lobes plus ou 
moins profondement lacinies, presque dichotomes. Fructi- 
fication inconnue." It is unnecessary to give any historical 
sketch of the various opinions expressed by later writers on 
the nature of this characteristic plant, but we may at least 
point out, that it has been held by certain authors that the 
plant regarded by Brongniart and others as an autonomous 
genus, was in all probability a particular form of calamitean 
branch. Stur was one of those who held this view, and in 

1 f ' Science Progress," vol. i., p. 54, and vol. i\\, p. 261. 

2 Williamson and Scott. 3 Broni:r.iart (1). PI. xiii., fig. 8. 
4 Brongniart (2), p. 6S. 



his great work on Calamites, several specimens are figured 
and described as evidence of the calamitean nature of 
Sphenophyllum. A restoration of Calamites with spheno- 
pylloid and other branches, given in his monograph, serves 
to illustrate this view. 1 More recent investigation has, 
however, conclusively proved that Brongniart's original 
definition holds good. There can no longer be any doubt 
that Sphenophyllum is a very well-defined generic type 
holding a somewhat isolated position in the plant king- 

From structureless casts and impressions, we learn that 
the genus is characterised by a comparatively slender 
articulated stem bearing a series of superposed whorls of 
leaves. The number of leaves in each verticil is always 
some multiple of three, frequently six, or it may be nine, 
twelve, eighteen, or more at each node. The leaves 
have usually a wedge-shaped form, and the lamina is 
traversed by dichotomously branched veins. In older forms, 
again, the leaves are much narrower, and each segment in a 
whorl has a single median vein. The narrow-leaved species, 
such as Sphenophyllum myriophyllum Crep., etc., 2 cannot 
always be readily distinguished from the well-known Astero- 
'bhyllites form of foliage; but as Zeiller 3 points out, a careful 
attention to the general habit of the plant, and the presence 
of bifui cations in the leaves, should enable us to separate 
these two generic forms. Another feature worthy of 
note is the hetrophylly occasionally exhibited by this 
genus. 4 The occurrence on the same plant of broad and 
finely dissected leaves, naturally suggested to some authors 5 
the idea of an aquatic plant ; but the histological features 
are not such as are usually associated with water plants. 

Examples of Sphenophyllum met with in English Coal- 
Measures do not, as a rule, attain any considerable length. 
By far the longest stem which has come under my notice is 
one in the Geological Survey Museum in Vienna ; in 
this specimen there is an axis 4 mm. in breadth with a 
length of 85 cm., giving off a slender branch 61 cm. in 

1 Stur, p. 69. 2 Zeiller (1), pi. lxii. 3 Zeiller (2), p. 674. 

4 Schenk, pi. xliv., fig. 1 ; Seward, p. 3, fig. 1. Etc. 5 E.g., Newberry. 


length. Occasionally long and narrow strobili are found 
attached to the vegetative branches ; in external appearance 
they resemble to some extent the corresponding structures 
in calamitean plants, but a closer inspection at once reveals 
a very distinct individuality for this type of strobilus. 

In Williamson and Scott's work three specific forms of 
Sphenophyllum are dealt with. We may first of all give a 
short description of the general type of structure char- 
acteristic of the genus, and afterwards attempt a diagnosis 
of the specific characters. 

Primary Structure of the Stem. — Traversing the young- 
stem there is a single vascular cylinder or stele, consisting 
of a triarch and centripetally developed axial strand of 
xylem. A transverse section of such a stem shows in 
the centre a triangular group of reticulate, scalariform, 
and spiral tracheids, the latter having a smaller diameter 
than the others, and constituting the three protoxylem 
groups at the prominent angles of the solid vascular axis. 
It is a fact of considerable interest, that we have in this 
primary structure an arrangement and manner of develop- 
ment of the tracheids which a student of Botany is always 
taught to regard as characteristic of root rather than stem 
structures. External to the xylem there is occasionally 
preserved a thin-walled phloem tissue, and beyond this 
may be recognised the pericycle or limit of the stele. 
Passing beyond the central cylinder we have a thicker 
walled cortex, of which the outermost layer or epidermis 
has not been clearly preserved. 

Secondary Structure. — On examining a series of trans- 
verse sections of stems in different stages of secondary 
growth, we find that the triangular group of primary 
tracheids becomes gradually surrounded by radially dis- 
posed rows of large elements, forming in older stems a 
considerable thickness of secondary wood, in which, as a 
rule, there is a striking uniformity in the diameter of the 
tracheae. Smaller xylem elements occasionally occur, but, 
as in the majority of Coal-period plants, there are no definite 
rings of growth. The development of secondary xylem, 
beginning in the interfascicular region, that is, in the broad 


bays of the primary wood, soon extends to the fascicular 
regions, and thus completely encloses the axial strand. 
The amount of secondary wood naturally varies con- 
siderably in different sections, the tracheids in a single 
radial row varying from one to thirty-seven in number. 
The medullary rays either extend as continuous lines of 
parenchyma through the whole thickness of the wood, or 
occur in the form of cell groups at the angles of the tracheids ; 
in the latter case the apparently isolated clusters of paren- 
chyma are united by connecting cells stretching across the 
radial walls of the reticulately pitted tracheids. Owing to 
the smaller diameter of the fascicular tracheae, the secondary 
xylem exhibits a fairly obvious division into six groups, 
three broader masses of interfascicular tracheids, alter- 
nating with three smaller groups of radial rows of fascicular 
tracheids, tapering towards the protoxylem angles of the 
primary xylem. 

The formation of periderm is another characteristic 
feature in the secondary growth of a Sphenophyllum stem. 
A phellogen or cork cambium appears to arise in the 
pericycle, and at a later stage the phloem parenchyma 
takes part in the development of cork tissue. It is often 
a matter of some difficulty to distinguish between the true 
phloem and the internal periderm. The latter consists of 
short cells in regular series, the former being made up of 
much longer elements, which may possibly be sieve-tubes. 

Leaves. — The most perfect example of a petrified leaf of 
Sphenophyllum so far described, is one figured by Renault. 1 
In transverse section the lamina is seen to be composed of 
thin-walled loose parenchyma, with small groups of trac- 
heids marking the position of the veins. The epidermal 
layer on the upper and under surface consists of fairly 
thick-walled cells, with indications of stomata. The most 
distinctly preserved stoma has, however, been figured by 
Solms-Laubach 2 from the epidermis of one of the leaf 
segments of a strobilus ; in this there are two narrow guard 
cells with two larger subsidiary cells. 

1 Renault (i), pi. ix., fig. 6; and (2), pi. xvi., fig. 1. 

2 Solms-Laubach, pi. x., fig. 9. 


Root. — As regards the roots of this genus we have but 
little information. Renault 1 has figured a small silicified ex- 
ample from Autun, with a diameter of 2 mm. In the 
centre there appears to be a diarch primary xylem bundle, 
surrounded by concentric rows of reticulately pitted trac- 
heicls. It is possible that two specimens figured by Felix, 2 
may representadventitiousrootsbeinggiven off from a.Sphc/10- 
phyllum stem. He speaks of them as examples of lateral 
branching, but their precise nature is, by no means, very 
easy to determine. 

Fructification. — The fructification of Sphenophyllum as 
first described by Williamson and Zeiller, 3 may be 
thus defined : — An axis traversed by a triangular strand of 
primary xylem tracheids, bearing at intervals of 1 "5 to 
2 - 5 mm. similar leaf verticils consisting of a number of 
linear lanceolate segments, fused in their lower portions 
into an open funnel-shaped disc. The numerous sporangia 
occur in 2 to 4 concentric circles on the upper surface of 
each disc, in radial sections of a cone presenting the appear- 
ance of a row of 2 to 4 sporangia between each whorl of 
bracts. Each sporangium is attached to a slender stalk 
springing from the upper surface of the leaf disc, and 
terminates in a hooked tip facing the axis of the strobilus, 
thus resembling the attachment of an anatropous ovule to 
its funicle. Each sporangiophore possesses a strand contain- 
ing a few xylem tracheids. At the point where the stalk or 
sporangiophore passes into the sporangium, the epidermal 
cells have thicker walls, and appear to represent an annulus, 
the sporangia dehiscing by a longitudinal slit on the side 
away from the stalk. The sporangia are isoporous, and the 
spores have a reticulately marked outer membrane. 

In a recent paper by Count Solms-Laubach 4 an exceed- 
ingly interesting addition is made to our knowledge of the 
Sphenophyllum strobilus. While confirming in the main 
the results arrived at by Zeiller, Williamson, and Scott, he 
describes a new type of fructification from the neighbour- 

1 Renault (1), pi. viii., fig. 5 ; and (2), pi. xv., fig. 6. 

2 Felix, pi. vi., figs. 2 and 7. 

3 See " Science Progress," vol. i., p. 54. 4 Solms-Laubach. 



hood of Cracow. In this species, Bowmanites Romeri, Solms, 
the sphorangiophores springing from the upper surface of 
each whorl of bracts, bear at the apex two sporangia instead 
of one as in previously known forms. Between each pair 
of sporophyll verticils there are at least three whorls of 
sporangia, the sporangia are almost sessile, and attached to 
short sporangiophes in the same manner as in the fructifica- 
tions already described. Each sporangiophore bifurcates 
towards the distal end, and the sporangia are attached to the 
diverging forks in much the samemannerastheovulesof^w?^ 
and Encephalartos are suspended from their carpophylls. 
As regards the nature of the spores and the annulus-like 
cells of the sporangial stalks, Bowmanites Romeri agrees 
closely with the other forms. As happens so frequently in 
pala^obotanical research, we are able to examine in detail 
the characters of an isolated member of an individual plant, 
without knowing anything of the other parts of the same 
species. In the present instance we are ignorant of the 
nature of the leaves which were borne by the stem to which 
Bowmanites Romeri was attached. There can, however, 
be little or no doubt that we have to do with a Spkeno- 
pkyllum strobilus, differing in an important respect from the 
ordinary type. The plants included in the genus Catamites 
are known to have possessed cones of more than one type 
of structure ; and it would appear that this was also the 
case with Sphenophyllum. When our data are more com- 
plete it may be possible to institute new generic terms for 
plants which are now assigned to these somewhat compre- 
hensive genera, but for the present it is better to err on 
the side of too wide a meaning for generic terms, than to 
attempt to found new genera on insufficient evidence. 

In addition to a full account of Bowmanites Romeri, 
Solms discusses at some length another sphenophylloid 
strobilus originally described by Weiss as Bowmanites 
Germanicus \ and suggests that this species as well as that 
described by Binney under the name of B. Camdrensis 2 

1 Weiss, PI. xxi., fig. 12. Solms-Laubach, PI. ix., fig. 7. 

2 Binney, PI. xii., figs. 1-3. 


may be identical with Bowmanites Dawsoni of William- 

It remains for us to consider the probable systematic 
position of this genus. It is undoubtedly a Vascular 
Cryptogam, characterised like so many other Palaeozoic 
representatives of the group by a considerable development 
of secondary xylem and phloem. 

Zeiller has expressed the opinion that Spkenophyllum 
should be included in the Filicince, and in the neighbour- 
hood of the ferns ; he institutes a comparison with Mar- 
siliacecE and Ophioglossece. The French author draws 
special attention to the distinct resemblance between the 
sporangiophore of Spkenophyllum and the sporocarp stalk 
in Marsilia. Subsequent writers have very properly 
pointed out that we cannot well make use of this super- 
ficial resemblance, in attempting to discover characters of 
real morphological importance. The single sporangium of 
Spkenophyllum differs in an important degree from the 
elaborate sporocarp or highly specialised foliar structure of 
Ma?-silia. The fossil genus is no doubt eusporangiate, 
and in that respect comparable with Ophioglossum, but the 
fertile spike of the latter differs widely from the sporangia 
and sporangiophores of the former. Potonies * comparison 
of Spkenophyllum with Salvinia does not render any 
material assistance to our endeavours to assign the fossil 
form to its true position. " We must be content for the 
present to leave this remarkable genus in its isolated 
position, in the hope that the extensive knowledge of its 
organisation which we now possess may in the future afford 
an adequate basis for comparison, when additional forms of 
Palaeozoic Cryptogams shall have been brought to light." 2 

This conclusion arrived at by Williamson and Scott, 
and accepted by Solms-Laubach, may perhaps be best 
realised by making use of the term SphenophyllecE as a 
class designation. This has been done by Schenk in 
Zittel's Handbtich der P alee onto logic? and is the course 
followed in a recent paper by Kidston 4 . 

1 Potonie. 2 Williamson and Scott, p. 946. 

3 Zittel. 4 Kidston. 




Class — Sphenophylle^e. 

Genus Sphenophyllum. — Brongniart 1828 (Sftkeno- 
pkyllites, Bronguiart, 1822). Stems comparatively slender 
( 1 *5 to 1 5 mm. ?), articulated, usually somewhat swollen at the 
nodes, and marked by more or less distinct ribs and grooves 
which do not alternate at the nodes, occasionally a single 
branch given off at a node. Leaves in verticils, usually the 
leaves of each whorl are equal in size, but may be unequal, 1 
in multiples of 3, 6, 9, 12, 18 or more. The leaves of 
successive whorls superposed, not alternate ; varying in 
form from cuneate, with narrow base and multinerved 
lamina having an entire or toothed anterior margin, to 
narrow linear uninerved forms, or with a deeply dissected 
lamina having dichotomously branched segments. 

Stem monostelic, with a triarch triangular strand of 
centripetally developed primary xylem, consisting of 
reticulate, scalariform and spiral tracheae ; the protoxylem 
elements being situated at the blunt corners of the xylem 
strand, from the angles are given off the foliar bundles, either 
one or two from each angle. 

Secondary xylem consists of radially disposed reticulately 
pitted tracheae, developed from a cambium layer. Phloem of 
thin walled tissue including sieve-pitted tube-like elements 
and phloem parenchyma. Both xylem and phloem traversed 
by medullary rays of parenchymatous cells. Cortex largely 
composed of fairly thick walled cells ; and in older stems 
cut off by the development of deep-seated periderm. 

Fructification in the form of long and narrow strobili, in 
some cases reaching a length of 12 cm., and a diameter of 
14 mm. A slender axis bearing whorls of numerous linear 
lanceolate bracts fused basally into a coherent funnel-shaped 
disc, bearing on its upper surface sporangiophores 2 and 

1 Zeiller (2), p. 675. 

2 The strobilus of S. trichomatosum, Stur, figured by Kidston, is de- 
scribed as having sessile sporangia. On this point see Williamson and 
Scott, p. 942. An examination of Kidston's specimen certainly conveys, 
the idea of sporangia without stalks, but the evidence is not conclusive. 


sporangia. Isosporous, possibly in some forms hetero- 
sporous. 1 

Sphenophyllum plurifoliatum. Williamson and Scott, 
Phil. Trans., vol. clxxxv., p. 920, pis. lxxv., Ixxxiii., 1894. 

Aster op kyllites Sphenopliylloides. Will. Phil. Trans., 
vol. i., p. 41, pis. i.-iv., 1874. [Type specimens from the 
Coal- Measures of Oldham in the Williamson Collection, 
British Museum.] 

Many linear leaves in each whorl ( 1 8 to 24 ?). Surface of 
young stems marked by three longitudinal grooves. 
Medullary rays in the form of groups of parenchymals 
cells in the spaces between the truncated angles of the 
secondary tracheae ; the groups connected laterally by 
means of radially elongated cells. Continuous rows of 
medullary ray cells rare. Deep seated periderm. 

S. insigne. (Williamson). Phil. Trans., vol. clxiv., 
p. 41, 1874, and Williamson and Scott. Phil. Trans., vol. 
clxxxv., p. 926, pis. lxxvi., Ixxxiii., lxxxiv., lxxxv., 1894. 

Aster op hyllites insignis. Williamson. Mem and Proc, 
Manchester Lit. and Phil. Society, vol. iv. [4], p. 13, 1891. 
[Type specimens from the Carboniferous beds of Burntis- 
land ; in the Williamson collection, British Museum.] 

Leaves probably not more than six in each whorl. 
Cortex grooved in young stems. Tracheae of primary 
xylem smaller in diameter than in .S. plurifoliatium. 

S. plurifoliatum. — Longitudinal canal at each angle of 
the primary xylem strand ; spiral tracheae more numerous 
than in the preceding species. Outer cortex of thinner 
walled cells than in .S*. plurifoliatum. Tracheae of secon- 
dary xylem with scalariform markings on radial walls. 
Medullary rays of regular rows, of one to two cells in 
breadth, extending through the entire thickness of the 
xylem. Phloem contains wide sieve- tube-like elements. 
Deep-seated periderm. 

In describing the fructification of Sphenophyllum, Wil- 

1 Kidston, in his definition of Sphenophyllum, speaks of it as hetero- 
sporous. The heterosporous example described by Renault is, however, 
extremely doubtful, and as yet we have no actual proof of the heterospory 
of this genus (see Kidston, p. 58 ; also Williamson and Scott, p. 942). 


liamson and Scott adopt the generic name of Sphenophyllum, 
while Count Solms prefers Binney's term Bowmanites. The 
question of terminology in paleobotany is often a difficult 
one. When we have very definite evidence that a cone 
belongs to a certain genus, it would appear the obvious 
course to speak of both under the same generic name. On 
the other hand there is something to be said in favour of 
retaining a special term for detached strobili, which cannot 
be certainly referred to their respective vegetative stems. 
In the case of Sphenophyllum Dawsoni (Will.) it may be, as 
suggested by Zeiller, the strobilus of S. cuneifolium (Sternb.); 
as our knowledge increases, detached cones must frequently 
be referred to certain specific forms of stems, and the con- 
fusion would probably be lessened if a distinct generic name 
were in the first instance assigned to isolated cones. The 
use of distinctive names for the fructification of genera has 
been found convenient in the case of Lepidodendron, Sigil- 
laria, and Calamites (Lepidostrobus, Sigillariostrobus, and 
C alamo stachys). Such names suggest the strobili of the 
different genera, and in looking through a list of species 
one recognises at a glance those which stand for repro- 
ductive structures. Solms does not adopt the generic 
designation of the fructification of Sphenophyllum corre- 
sponding to C alamo stachys, as he considers such a term as 
Sphenophyllostachys too long and inconvenient. In coining 
new names sesquipedalian words should, as a rule, be 
avoided, but in discarding the genus Sphenophyllostachys 
one is departing from a recognised and convenient custom 
for a reason which hardly seems adequate. Bowmanites is 
the older name, but now that its true position is known, it 
should be replaced by a term which expresses the fact of 
its connection with Sphenophyllum. I would suggest, 
therefore, that the name Sphenophyllostachys be adopted 
for the strobili of Sphenophyllum. 

Sphenophyllostachys Dawsoni (Will.). {Mem. Man- 
chester Lit. and Phil. Soc, vol v., p. 28, pis. 1-3, 187 1.) 

Volkmannia Dawsoni. Ibid. 

Bowmanites Dawsoni. Weiss. Steinkohlen Calamarien, 
ii., p. 200, 1884. 


Slender axis bearing alternating whorls of bracts (14 to 
20), cohering basably and free distally as long linear segments 
extending- upwards through about six internodes. A single 
verticil of long and slender sporangiophores on each whorl 
of bracts. The sporangiophores bend inwards at the apex 
and bear single sporangia. Isosporous. Spores with spinus 
outgrowths. Probably the strobilus of Sphenophyllum 
cuneifolium (Sternb). 

Sphenophyllostachys RoMERi(Solms-Laubach). Jahrb. 
Geo/. Reichs. Wien, Bd. 45, Heft. 2, p. 225, Pis. ix. and x., 

i895- _ 

Axis and whorls of bracts similar to those of 5. Dazvsoni, 

except that in each verticil of bracts the free linear seg- 
ments extend nearer to the strobilus axis. More than one 
whorl of sporangiophores on each whorl of bracts, probably 
three. Each sporangiophore forked distally, and bearing 
a sporangium on the inwardly bent tip of each diverging 
branch. Isosporous. Spores similar to those of 5". 

Genus Trizygia. Royle. Botany and Nat. Hist. 
Himalayan Mts., p. 431, 1834. 

This generic name was proposed in 1834 for a genus of 
plants occurring in the Glossopteris flora of India. 1 Little 
is known as to its real affinities or structure, but Zeiller 2 
has recently pointed out the doubtful generic value of its 
characters, and he regards it as most probably a form of 
Sphenophyllum. The slender stem bears verticils of wedge- 
shaped leaves in three pairs at each node, the anterior pair 
being smaller than the two lateral pairs. 


Binney, E. W. Observations on the Structure of Fossil Plants in 
the Carboniferous Strata. Palceontological Society, pt. ii., 1871. 

BRONGNIART, A. (i). Sur la classification et la distribution des 
vegctaux fossiles en general, 1822. 

Brongniart, A. (2). Prodrome d'une histoire des vegetaux fossiles, 
Paris, 1828. 

1 For figures see Feistmantel, pis. xi. A. and xii. A. 2 Zeiller (2). 


Feistmantel, O. The Fossil Flora of the Lower Gondvvanas. 

II. The Flora of the Damuda and Panchet Divisions. Mem. 

Geo/. Surv. India, vol. hi., Calcutta, 1881. 
FELIX, J. Untersuchungen iiber den inneren Bau Westfalischer 

Carbon-Pflanzen. K. Preuss. Geo/. Landesanst., p. 153, 1886. 
KlDSTON, R. On the Fructification of Sp/ienopJiy/lum trichomatosum, 

Stur, from the Yorkshire Coal Field. Proc. R. Phys. Soc., 

Edinburg/i, vol. xi., p. 56, 1890-91. 
NEWBERRY, J. S. The Genus Sphenophy/lum. Journ. Cincinnati 

Nai. Hist. Soc., p. 212, 1891. 
POTONIE, H. Ueber die Stellung der Sphenophyllaceen im System. 

Bericht. dentsch. bot. Gese//., Band, xii., Heft 4, p. 97, 1894. 
RENAULT, B. (i). Nouvelles recherches sur la structure des 

SphenopJiyllum et sur leurs affinites botaniques. Ann. Sci. 

Nat. Bot. [6], vol. iv., p. 277, 1876. 
Renault, B. (2). Cours de botanique fossi/e, vol. ii., 1882. 
Schenk, A. In Richthofen's China, vol. iv., Berlin, 1883. 
Seward, A. C. Sp/ienopJiyUum as a Branch of Asterophy/lites. 

Mancliester Lit. and Phi/. Soc, p. 1, 1890. 
Solms-Laubach, Graf zu. Bowmanites Romeri, eine neue 

Sp/tenopJiylluni Fructification. Ja/irb. k. k. Geo/. Reichsanust. 

Wien, vol. xlv., Heft I, p. 225, 1895. 
Stur, D. Die Carbon-Flora der Schatzlarer Schichten. Abth. II. 

Die Calamarien. K. k. Geo/. Reichs., vol. xi.; Abth. II., Wien, 

WEISS, C. E. Beitrage zur fossilen Flora, III. Steinkohlen- 

Calamarien, II. Abth. Geo/. Specia/karte Preuss. Thiiring- 

Staaten, Band v., Heft. 2, Berlin, 1884. 
WILLIAMSON, W. C, and Scott, D. H. Further Observations on 

the Organisation of the Fossil Plants of the Coal-Measures. 

Part I. Catamites, Ca/amostachys, and Sphenophy//um. Phi/. 

Trans., vol. clxxxv. B., p. 863, 1894. 
Zeiller, R. (i). Bassin houiller de Valenciennes. Etudes de gites 

Minkraux, Paris, 1886. 
Zeiller, R. (2). Sur la Valeur du Genre Trizygia. Bu/L Soc. 

Geo/. France [3], vol. xix., p. 673, 1891. 
Zittel, K. A. VON. Handbuch der Pa/a?onto/ogie. Abth. II. 

Palaeophytologie. Schimper & Schenk, 1890. 

A. C. Seward. 


THE geology of the Spanish Peninsula is imperfectly 
understood; but it is not without a special interest of 
its own. It is here, if anywhere in Europe, that we should 
expect to find among the more ancient faunas some indica- 
tions ot a warmer temperature than prevailed farther north, 
or of some other difference due to difference of latitude. 
It is the only part of Southern Europe where there is a really 
extensive development of the Lower Palaeozoic rocks ; but 
unfortunately these are still almost unknown. 

Of late years the re-organisation of the geological surveys 
of Spain and Portugal has led to a great increase in our 
knowledge of those countries, and the recent appearance of 
a new part of the Comnmnica(pes da Direc(ao dos Trabalhos 
geologicos de Portugal affords a good opportunity of recapitu- 
lating what has been accomplished in that country. 

It is impossible to look at this and the other publications of 
the Portuguese Survey without a word of praise for the beauty 
of the plates with which they are illustrated, and the admir- 
able way in which they are printed. It is a painful reflection 
to an English geologist that the inimitable work of our own 
Geological Survey should be presented to the world in a 
style so far inferior ; and that the enlightened Government 
of a great empire should in this respect be so far behind 
that of a small and not very wealthy country like Portugal. 

We may, however, be allowed to express our regret that 
so few of the memoirs of the Portuguese Survey are 
accompanied by maps, for without a map it is extremely diffi- 
cult to follow the text of a stratigraphical paper; and without 
making a map it is, or should be, almost impossible for the 
worker himself to be sure that his views are correct. We 
regret too the long delay in the publication of a new edition 
of the general geological map of Portugal. Delgado is 
twenty years old, and although later information is else- 
where available, it is surely time that the Geological Survey 


should take upon itself the production of a map more in 
accordance with modern needs. 

By far the greater part of Portugal is occupied by ancient 
rocks of Archaean and Palaeozoic age, and by eruptive 
masses which probably belong to various periods. All the 
higher mountains are formed of such rocks ; and it is only 
in the plain of the Tagus and along the coasts that any later 
beds are to be found. The most extensive area of Mesozoic 
rocks forms a broken triangle with its base parallel to the 
Tagus between Lisbon and Torres Novas, and its apex at 
Oporto. Mesozoic rocks also occupy a narrow strip of 
country along the southern shores of Portugal in the 
province of Algarve. They are found too in the Serra da 
Arrabida, which forms the prominent cape south of Lisbon ; 
and at Sao Thiago de Cacem and Cabo de Sines farther 
south. The largest area of Tertiary deposits is that which 
forms the plain watered by the Tagus and its tributaries. 

" Azoic " Rocks. — The so called Azoic rocks, in which 
no fossils have hitherto been discovered and which are pre- 
sumed to be older than the Cambrian, are best developed 
east of the Tertiary basin of the Tagus in the province of 
Alemtejo. But it cannot be said that their age has been 
determined with certainty, and the supposed absence of 
of fossils may be due to imperfect examination. 

Lower Palceozoic. — In spite of the extensive area occu- 
pied by schists and other rocks of supposed Palaeozoic age 
undoubted Cambrian fossils have been found at only a 
single locality in Portugal. So long ago as 1876, between 
the "Azoic" rocks of Alemtejo and the Lower Carboniferous 
of the borders of Algarve, Delgado recognised a series of 
beds which he believed to be distinct from both ; and in 
this series, near the mines of San Domingos, were found 
Nereites and other forms which are usually believed to be 
tracks of animals. They are quite insufficient to determine 
the age of the beds, and it was chiefly from a lithological 
resemblance to certain rocks in the North of Portugal that 
Delgado referred them to the Cambrian (15). 

More recently (21), however, trilobites have been dis- 
covered near Villa Boim, some 10 km. west of Elvas ; and 


these trilobites appear to belong to the characteristic Cam- 
brian families Olcnidcs and Conoccphalidce. Delgado, indeed, 
compares them with the genera Liostracus and Lcptoplas- 
ttis ; but unfortunately no figures have yet been published, 
and all that the descriptions enable us to say is that they 
probably belong to the Olenus group. This discovery is of 
oreat interest, for at one time it was believed that the 
Olenus fauna was absent in Southern Europe. Recently, 
however, it has been found also in Sardinia. 

The Ordovician and Silurian rocks are much better de- 
veloped than the Cambrian, or at least they have been far 
longer known and have yielded fossils much more abun- 
dantly. One of the best known localities is that of Val- 
longo, 10 km. E.N.E. of Oporto, where Sharpe obtained 
a number of Ordovician fossils which were described by 
himself and others (41). Recently, Delgado has published 
a new list of the forms from this neighbourhood, and he re- 
cognises three distinct horizons (19). But there is some 
confusion in the identification either of the horizons or of 
the fossils ; for from the third horizon he records, for 
example, both the Lower Ordovician form Acidaspis Buchi 
and the Silurian species, Phacops Doivningice. Most of the 
Vallongo specimens are certainly Ordovician, and among 
them are Placoparia, Calymene, Tristani and others, charac- 
teristic of the Anoers slates of France. 

At Bussaco, some 20 km. north of Coimbra, Silurian 
fossils, as well as Ordovician, are found in some abundance. 
The Ordovician beds consist of a lower division of quart- 
zites, sandstones, black shales and limestones, with Caly- 
mene Tristani, Placoparia Zippci, etc. ; and these are suc- 
ceeded by ochreous argillaceous rocks with Phacops Dujar- 
diui, etc. The Silurian is represented by blue shales and 
argillaceous schists with graptolites, Cardiola interrupts and 
the thin-shelled Orthoceratites which Forbes called Creseis 
(38). Such forms are characteristic of our Lower Ludlow, 
and to a certain extent of our Wenlock beds. 

South of the Tagus, in the neighbourhood of the grani- 
tic mass of Portalegre, Delgado has reported the presence 
of Ordovician beds (15). They commence with a series of 


quartzites containing numerous " bilobites," similar to those 
which in the North of Portugal, and in parts of France, are 
found at the base of the Ordovician system. It is unneces- 
sary here to enter into the controversy concerning these 
forms. Nathorst has given strong reasons for believing 
them to be the tracks of animals ; but Delgado strongly 
opposes this view and maintains them to be algae (16). 

In the neighbourhood of Portalegre there is found also 
a small patch of schists containing Monograptus and some 
casts of bivalves (15). The relations of these to the sur- 
rounding beds are unknown ; but if the graptolites are 
correctly referred to the genus Monograptus, they must 
certainly belong to the Silurian. 

So far then as they are known, the Lower Palaeozoic 
rocks of Portugal do not favour very strongly the view that 
there was any very marked difference in Older Palaeozoic 
times between the faunas of Northern and Southern Europe. 
Nevertheless, P/acoparia, Calymene Tristci7ii, and Acidas- 
pis Buchi, which are characteristic fossils in France and the 
Spanish Peninsula, are by no means common in Britain, 
although they have been found there. In short, we have 
no sufficient data as yet to show how far the fauna of 
Southern Europe resembled or differed from that of the 

Upper Palceozoic. — There is only one locality in the 
whole of Portugal where the Devonian has yet been 
recognised, and this is near the Ordovician quartzites of 
Portalegre. A band of schists was discovered by Delgado 
containing Phacops latifrons, Cryp/urus, and broad-winged 
Spirifers (15). 

The Lower Carboniferous on the other hand occupies a 
wide area in the South of Portugal, where it forms the 
greater part of the hilly region on the northern borders of 
the province of Algarve. Like all the other Palaeozoic 
rocks of Portugal, they have never been studied in detail, 
but they consist of schists and grauwackes, without either 
quartzites or limestones, and they contain Posidonomya 
(like Becker i and Pargai), and Goniatites [cf. crenistrid) 
{15). Hence it appears that the Lower Carboniferous 


belongs to the "Culm" facies so widely developed in 
Central Europe. 

The Upper Carboniferous is very restricted in extent, 
and its distribution bears no relation whatever to that of 
the Lower Carboniferous. From the character of the de- 
posits, and the mode of their occurrence, there can be little 
doubt that the Upper Carboniferous of Portugal was laid 
down in comparatively small basins not unlike those of 
the Central Plateau of France. It invariably rests uncon- 
formably upon very much older beds, and consists very 
largely of coarse conglomerates. 

The most extensive area is met with in the North of 
Portugal, where the coal measures form a band stretching 
from the sea-coast at Espozendo (North of Oporto) in a 
S.S.E. direction across the Douro as far as Pijao in the 
province of Beira (22). The coal of this band near 
Vallongo was taken by Sharpe to be of Silurian age 


Farther north there is another band some 22 km. 
long and 700 m. wide in the neighbourhood of Bussaco (38, 
26) ; and lastly, South of the Tagus there is a very small 
patch at Moinho d'Ordem near Alcacer do Sal (28). 

The fossil plants from these three basins have been 
described by several writers, and according to Gomes (22) 
they indicate that the deposits of all three are of the same 
age, viz., that of our Coal Measures. But Wenceslau de 
Lima has recently revised the flora of Bussaco, and from 
various considerations, and especially from the presence of 
Walchia and Callipteris, he has been led to conclude that 
the coal bearing deposits of Bussaco belong to the Roth- 
liegende, or to the passage beds between Carboniferous 
and Permian (26). He believes, however, that the coal of 
the other basins is of somewhat earlier date. It is remark- 
able that in the Bussaco beds a crustacean has been found 
which W. de Lima refers to the genus Eurypterus 

(27) - 

Trias. — The Palaeozoic rocks of Portugal are uncom- 
formably overlain by a series of red and white sandstones 
and conglomerates, to which Choffat has given the name of 


"gres de Silves ". North of the Tagus these sandstones 
form the eastern border of the Mesozoic area, stretching in 
a narrow band from Aveiro nearly to the town of Thomar. 
South of the Tagus the " gres de Silves " is met with at Sao 
Thiago de Cacem, at Carrapateira (N. of Cape St. Vincent), 
and again as a narrow band resting upon the Palaeozoic 
rocks in the littoral region of the province of Algarve (9). 

North of the Tagus the sandstones begin to alternate in 
the upper part of the series with dolomitic and argillaceous 
limestones, and these in turn are surmounted by dolomites 
without sandstone belonging to the Sinemurien (40). In 
the lower or sandy part of the series there are several beds 
which contain remains of plants. These have been exa- 
mined by both Heer (23) and Saporta (40) and seem to 
indicate a Rhaetic or Infraliassic aore. 

The calcareous beds which occur higher up in the series 
are called the " beds of Pereiros ". They often contain 
marine fossils, for the most part gastropods and lamelli- 
branchs, which are believed by Choffat to belong to 
the Infralias. 

In Algarve the general succession is very similar. The 
red sandstones of the lower part of the series have yielded 
no fossils. The dolomites of the upper part contain marine 
forms ; and as in the north, the dolomites gradually increase 
at the expense of the sandstones. They are overlaid by 
marls, spotted with white, which often contain gypsum, but 
no fossils (9). 

Jurassic. — Jurassic rocks are found in four separate areas, 
all of which are in the neighbourhood of the coast-line. 
They extend, with some interruptions, from Aveiro in the 
north of Portugal, to Cintra. They occur also in the 
Serra da Arrabida, which forms the promontory south of 
the Tagus which we call Cape Espichel. They are found 
in the next important Cape to the south, the Cabo de Sines ; 
and lastly, they are extensively developed along the coast 
of the province of Algarve. 

The system is divided by the Portuguese geologists 
into three stages, corresponding with the three divisions 
adopted in Central Europe and named in ascending order,, 


the Lias, Dogger, and Malm. There is, unfortunately, no 
general account of the whole ; but Choffat has given us a 
brief description of the Jurassic of Algarve (9), and a 
fuller account of the Lias and Dogger (2) and of the lower 
part of the Malm (12) north of Lisbon. 

The Lias and Dogger are almost entirely marine, and 
correspond very closely with the contemporaneous beds 
of Central Europe. Many of the northern zones have 
been recognised in Portugal and further research no doubt 
will reveal others. It may be noted here that Choffat 
includes the Callovian in the Dogger (5, 8). 

The Malm, on the other hand, is a much less purely 
marine formation, and in places contains beds of lignite, 
which are sometimes, for example, at Cape Mondego, 
extensive enough to be worked. It differs considerably 
from that of Northern Europe, and is very variable in 
character. Everywhere, however, it may be divided into 
two stages, the Lusitanian below and the Neo-jurassic 

North of the Tagus the Lusitanian as it exists in 
the country of Torres Vedras has been described by